Skip to main content

2020 | Book

XV Mediterranean Conference on Medical and Biological Engineering and Computing – MEDICON 2019

Proceedings of MEDICON 2019, September 26-28, 2019, Coimbra, Portugal

Editors: Jorge Henriques, Nuno Neves, Prof. Dr. Paulo de Carvalho

Publisher: Springer International Publishing

Book Series : IFMBE Proceedings


About this book

This book gathers the proceedings of MEDICON 2019 – the XV Mediterranean Conference on Medical and Biological Engineering and Computing – which was held in September 26-28, 2019, in Coimbra, Portugal. A special emphasis has been given to practical findings, techniques and methods, aimed at fostering an effective patient empowerment, i.e. to position the patient at the heart of the health system and encourages them to be actively involved in managing their own healthcare needs. The book reports on research and development in electrical engineering, computing, data science and instrumentation, and on many topics at the interface between those disciplines. It provides academics and professionals with extensive knowledge on cutting-edge techniques and tools for detection, prevention, treatment and management of diseases. A special emphasis is given to effective advances, as well as new directions and challenges towards improving healthcare through holistic patient empowerment.

Table of Contents


Regular Sessions: Biomedical Signal Processing

Using Eye Tracking to Analyze Surgeons’ Cognitive Workload During an Advanced Laparoscopic Procedure

Introduction: Surgeons’ cognitive workload should be controlled during a surgical intervention for a successful and safe performance. Eye tracking technologies could be used for cognitive workload monitoring by analyzing the pupil size and blink frequency of surgeons. This work aims to study the surgeons’ cognitive workload watching an advanced laparoscopic video. Methods: 16 surgeons (5 experts, 11 novices) participated in the study watching a colectomy video consisting of eight steps with introductory titles. Surgeons’ gaze was recorded with eye tracking glasses while visualizing the video, from which pupil size and blink frequency were analyzed. Results: Pupil diameter of surgeons increased during the visualization of steps and decreased during the titles. Two specific steps of the intervention produced the highest pupil diameter of surgeons. When the pupil diameter increased the blink frequency decreased. Pupil diameter tended to decrease as the video is watched, which is mainly due to expert surgeons, who had a lower pupil diameter than novices. Conclusions: Eye tracking technologies allow monitoring the cognitive workload of surgeons in surgical procedures. Larger pupil size and shorter blink frequency means greater cognitive workload. Such metrics could be used to objectively label the difficult tasks within the surgical procedure. Surgical videos used for training of surgeons should be short, according to micro-learning, since cognitive workload decreases over time while visualizing them. Based on proposed metrics, eye tracking technologies could be used to distinguish the level of experience of surgeons, since cognitive workload is sensitive to the skill level of surgeons.

Juan Francisco Ortega-Morán, J. Blas Pagador, Vicente Luis-del-Campo, Juan Carlos Gómez-Blanco, Francisco M. Sánchez-Margallo
Application of Multivariate Spectral F Test for Somatosensory Evoked Response Detection

Somatosensory Evoked Potential (SEP) is an important tool for monitoring vascular and spine surgeries, and other clinical applications. However, morphological SEP identification is subjective. Then, statistical techniques, such as Local Spectral F Test (SFT), have been used for response detection. The multivariate extension of SFT employs more than one derivation and has been recently considered advantageous to identify response to visual stimulation. This work aims at evaluating the performance of Multivariate SFT (MSFT) applied to EEG signals from 40 volunteers during stimulation at 5 Hz and different numbers of derivations (N), comparing the detection rates (DR). Frequencies of interest fo1 = 15 Hz and fo2 = 100 Hz were used, as well as L = 6 neighbor components at the frequencies from 70 to 95 Hz and a 5%-significance level. The number of derivations varied from N = 1 to 6. The detection rates obtained using fo1 were higher than those with fo2, which corresponds to false positives, since no response is expected to occur at such frequency. For fo1, half of the volunteers exhibited a monotonic increase of DR as N was augmented. For other volunteers, an oscillatory pattern was noticed in DR as new derivations were added, suggesting that raising N do not necessarily lead to improvement in MSFT performance. For fo2, raising N caused an increase in the false-positives rate above significance level of 5%. This could be explained in part by the correlation among the employed derivations. Finally, MSFT showed promising results at SEP identifying.

Karina Miranda Boson, Antonio Mauricio Ferreira Leite Miranda de Sá, Danilo Barbosa Melges
Spatial Cross-Correlation to Determine Atrial Fibrillation Recurrence After Ablation

Atrial fibrillation (AF) is already the most frequent arrhythmia in the clinical practice. Pulmonary vein ablation has emerged as a treatment that is able to make disappear the arrhythmia, nevertheless up to half of ablated AF patients suffer recurrences within the first year. In this study multiple simultaneous intra-atrial bipolar recordings were located at three locations, pulmonary veins and the right and the left atrium before the ablation procedure. Signal organization parameters were determined using cross-correlation analysis along the three main atrial regions. The results showed that before the procedure, patients with recurrence in the arrhythmia showed that left atrial activity was less correlated with other atrial regions compared with patients that maintained sinus rhythm. Moreover, time-lagged correlation coefficient between two time series showed higher delays in patients that experienced recurrence of AF. These results can be helpful for procedures designing to end AF.

Raquel Cervigón, Julián Pérez-Villacastín, Javier Moreno
Development of a Computer Simulator of the Visual N2 Event-Related Potential Component for the Study of Cognitive Processes

The importance of visual N2 Event-Related Potential (ERP) component in the study of cognitive processes lies in its interpretation as a measure of the allocation of visual attention to possible targets. Unfortunately, the N2 component has a small amplitude and the domain of validity of methods used for its estimation is difficult to assess if real data are considered. Here, we develop a computer simulator of ERP measurements emulating the variability of the visual N2 component and use the synthetic data to evaluate the performance of four popular literature ERP estimation methods. Results confirmed that a solid simulation framework could allow identifying a reliable method to detect small-amplitude ERP components and quantifying its accuracy.

Francesca Marturano, Sabrina Brigadoi, Mattia Doro, Roberto Dell’Acqua, Giovanni Sparacino
Automatic Segmentation of Ultrasonic Vocalizations in Rodents

Ultrasonic vocalizations studies in rodents have increasingly drawn researchers attention as it have been considered a powerful tool to understand the animals behavior and their interactions in different social and environmental contexts. This paper presents an entropy-based algorithm for accurate and robust segmentation of mouse ultrasonic calls. Instead of using the conventional energy-based features, the spectral entropy is developed to identify the audio segments accurately. The new approach for mice calls detection has been able to detect up to 97% of the vocalizations.

Diogo Pessoa, Lorena Petrella, Miguel Castelo-Branco, César Teixeira
PCG-Decompositor: A New Method for Fetal Phonocardiogram Filtering Based on Wavelet Transform Multi-level Decomposition

Fetal phonocardiography (FPCG) is a non-invasive acoustic recording of fetal heart sounds (fHS). The fHS auscultation plays an important diagnostic role in assessing fetal wellbeing. Typically, FPCG is a non-stationary signal corrupted by the presence of noise. Thus, high-amplitude noise makes detection of FPCG waveforms challenging. Thus, appropriate filtering procedures have to be applied in order to make FPCG clinically usable. In the recent years, Wavelet transformation (WT) filtering has been proposed. In particular, aim of this study is to propose a new method based on WT multi-level decomposition filtering: PCG-Decompositor. To this aim, PCG-Decompositor based on Coiflets mother Wavelet (4th order, 9 levels of decomposition) was applied to 119 real FPCG tracings, all available in Physionet. PCG-Decompositor is a dependent thresholding technique based on FPCG multi-level decomposition analysis. Performances of PCG-Decompositor are computed against soft-thresholding denoising technique (STDT) in terms of Root Mean Square Error (RMSE) and fetal heart rate (fHR). In terms of fHR, PCG-Decompositor and STDT are compared between themselves and also with the so-called annotations, given by the average fHR using a simultaneous cardiotocography analysis. Original signal to noise ratio (SNR) values ranged from 7.1 dB to 24.4 dB; after application of PCG-Decompositor, SNR increased significantly, ranging from 9.7 dB to 26.9 dB (P < 10−7). Moreover, PCG-Decompositor showed a lower dispersion than STDT (RMSE: 0.7 dB vs. 1.2 dB), introduced no FPCG signal delay and left fHR unaltered. Thus, PCG-Decompositor could be a suitable and robust technique to denoise FPCG signals.

Annachiara Strazza, Agnese Sbrollini, Marica Olivastrelli, Agnese Piersanti, Selene Tomassini, Ilaria Marcantoni, Micaela Morettini, Sandro Fioretti, Laura Burattini
Muscular Co-contraction Detection: A Wavelet Coherence Approach

Muscular co-contraction is defined as the activity of agonist and antagonist muscles around a joint, enhancing joint stability and balance. The quantitative assessment of muscle co-contractions would be meaningful for deepening the comprehension of this physiological mechanism. Thus, the purpose of this work is to quantify muscle co-contraction using a Wavelet transform-based coherence analysis of sEMG signal during straight walking. To this purpose, sEMG from tibialis anterior (TA) and gastrocnemius lateralis (GL) and basographic signals were acquired in five healthy subjects during walking. Basographic signals were analyzed to quantify foot-floor contact. sEMG signals were processed using Wavelet Transform Coherence (WTC) to identify muscular co-contractions. Daubechies (order 4 with 6 levels of decomposition) was chosen as mother wavelet. A denoising algorithm based on a soft thresholding was applied for removing noise from raw signals. Denoised signals were considered to achieve WTC function, a well-known localized statistical assessment of cross-correlation between signals. Thus, in this work, WTC cross-correlation could be considered to assess muscular co-contraction. This methodology applied to TA and GL signals was able to detect GL/TA co-contractions during hell-strike (0–10% of GC) phase and during P-phase (54.2–68.3% of GC), matching with literature. Moreover, WTC approach was able to provide also the frequency band of information content for muscle co-contractions: 32–65 Hz for H-phase co-contraction and 16–32 Hz for P-phase co-contraction. In conclusion, this study proposed WTC analysis as a reliable method to assess muscle co-contraction in time-frequency domain.

Annachiara Strazza, Federica Verdini, Andrea Tigrini, Stefano Cardarelli, Alessandro Mengarelli, Sandro Fioretti, Francesco Di Nardo
Calculation of Breath-by-Breath Oxygen Uptake in Asthmatic Patients by the “Independent Breath” Algorithm. Comparison with a Classical Approach

Several computation algorithms are available to determine gas exchange on a breath-by-breath basis, each designed on the basis of different theoretical backgrounds, including the newly proposed “Independent breath” algorithm. The new algorithm was tested on the respiratory signals acquired from 11 asthmatic patients and 20 well-matched healthy controls, comparing results also with those provided by a “classical” algorithm commonly applied by other laboratories. Oxygen, carbon dioxide fractions, and ventilatory flow were recorded at the mouth continuously over 26 min in all the volunteers at rest, during unloaded and moderate intensity cycling and subsequent recovery. Average oxygen uptake values calculated for the 4 steady-state conditions (over 3 min), as well as the corresponding standard deviations, were not significantly different between the two groups of subjects (MANOVA, Group effect, p = ns). Almost all the average oxygen uptake values provided by the two algorithms were overlapping, the large majority lying within 5% from the identity line. The corresponding standard deviations obtained for the “Independent breath” algorithm were lower than those obtained for the “classical” algorithm (MANOVA, Algorithm effect, p < 0.001), the slope of the regression line between them amounting to 0.672. In conclusion, because of its better precision, with similar accuracy, compared to the “classical” real-time breath-by-breath algorithm, the use of the “Independent breath” algorithm should be recommended, also in asthmatic patients.

Maria Pia Francescato, Miloš Ajčević, Valentina Cettolo, Mario Canciani, Agostino Accardo
Gait Phase Classification from Surface EMG Signals Using Neural Networks

Identification and classification of different gait phases is an essential requirement to temporally characterize muscular recruitment during human walking. The present study proposes a Deep-learning methodology for the classification of the two main gait phases (stance and swing), based on the interpretation of surface electromyographic (sEMG) signals alone. Three different Multi Layer Perceptron (MLP) models are tested to this aim. The present approach does not require specific features to be extracted from the signal, differently from previous studies. 12 healthy adult subjects are analyzed during walking over-ground at comfortable speed. sEMG signals from eight leg muscles are selected. Performance of classifiers is tested vs. gold standard, represented by basographic signals measured by means of three foot-switches. A 10-fold evaluation is computed to take into account the possible variability of the results. The direct comparison among the performances of the three different MLP models shows an average high accuracy over the population (around 95%) for all the models, independent from the increasing complexity. Moreover, the accuracy in each single subject does not fall below 92.6% (range of accuracy variability = 92.6–97.2%). This present study suggests that artificial neural networks may be a suitable tool for the automatic classification of gait phases from electromyographic signals, in overall walking tasks.

Christian Morbidoni, Lorenzo Principi, Guido Mascia, Annachiara Strazza, Federica Verdini, Alessandro Cucchiarelli, Francesco Di Nardo
Combining Objective Response Detectors Using Genetic Programming

Many Objective Response Detectors (ORD) have been proposed based on ratios extracted from statistical methods. This work proposes a new approach to automatically generate ORD techniques, based on the combination of the existing ones by genetic programming. In this first study of this kind, the best ORD functions obtained with this approach were about 4% more sensitive than the best original ORD. It is concluded that genetic programming applied to create ORD functions has a potential to find non-obvious functions with better performances than established alternatives.

Leonardo Bonato Felix, Quenaz Bezerra Soares, Antonio Mauricio Ferreira Leite Miranda de Sá, David Martin Simpson
Handwriting Kinematic Differences Between Copying and Dictation

Handwriting is a human activity that may be affected by the modality used as input of the information to be written, mainly copying or dictation. Many processes at different levels are involved to produce motor planning and graphomotor automation of handwriting. In order to quantify possible kinematic differences due to the influence of auditory or visual input modalities to these processes, three different tests were proposed to a sample of 101 young students and several kinematics parameters measuring handwriting characteristics were evaluated. The tests required to copy as accurate (CA) and as fast (CF) as possible an Italian sentence and to write the same sentence under dictation (DF). All parameters showed significant differences between each pair of the three tests. The best performance was obtained in the CF test followed by the DF and CA tests; in the latter the greater accuracy required to produce writing yielded lower velocity and automation as well as a longer motor planning time. On the other hand, the dictation response was more similar to that of CF than CA showing a larger planning time, probably due to a different time necessary to correctly identify the words to reproduce. The combination of the two tests could be useful to study the impairment of either visual or auditory input.

Silveri Giulia, Accardo Agostino
Bradycardia Assessment in Preterm Infants

Prematurity is a severe condition, usually correlated with critical outcomes. One of the major diseases in preterm infants is bradycardia, defined as the heart rate decreasing under 100 bpm for at least two heartbeats in duration. Usually, bradycardia is considered as a manifestation of immature cardiorespiratory control, but no studies investigated its nature in relation to the different clinical features of preterm infants. Thus, aim of this work is to assess the relation between bradycardia features and the main preterm infant clinical features, weight and gestational age. Ten preterm infants were considered, classified according with three criteria: the weight classification, the gestational age classification and the birth size assessment (that combined the two previous classifications). For each preterm infant, bradycardias are automatically identified and characterized in term of bradycardia features: amplitude, duration and area. Moreover, bradycardia events are classified according with their severity. Finally, bradycardia feature distributions of classes that belong to the same classification criterion were compared. Results seems suggesting that bradycardia features differences are more relevant in preterm infants with different weights than in those with different gestational age, contrary to what expected. Anyway, the best results in term of classification were obtained in the birth size assessment; thus, a combined approach that considers both weight and gestational age is preferable. Moreover, a combined evaluation of amplitude and duration for bradycardia characterization can better assess the severity of this arrhythmia and of the preterm infant clinical status.

Agnese Sbrollini, Martina Mancinelli, Ilaria Marcantoni, Micaela Morettini, Laura Burattini
To What Extent Does Heart Rate Alter the Cerebral Hemodynamic Patterns During Atrial Fibrillation?

Atrial fibrillation (AF), the most common cardiac arrhythmia leading to irregular and faster heartbeat, has been recently and independently associated to the risk of dementia. A constellation of hemodynamic mechanisms has been proposed to explain the possible link between the two pathologies. However, definitive data still miss, and it is unknown how heart rate (HR) influences the cerebral microcirculation. We propose a computational approach, based on a validated hemodynamic modeling, to compare the cerebral circulation during normal sinus rhythm (NSR) and AF at different HRs. AF is able to trigger a higher variability of the cerebral blood flow variables which grows towards the distal circulation. The alteration of the hemodynamic patterns, inducing the rupture of the signal periodicity and the consequent higher occurrence of extremely high/low values, increases with HR. Awaiting necessary clinical evidences, present findings highlight that a strict rate control strategy could be beneficial in terms of cognitive outcomes in patients with permanent AF.

Stefania Scarsoglio, Luca Ridolfi, Andrea Saglietto, Matteo Anselmino
Non-invasive Intrauterine Pressure Estimation Based on Nonlinear Parameters Computed from the Electrohysterogram

Monitoring uterine contractions is essential during pregnancy and labor to obtain information on time-to-delivery and maternal and fetal wellbeing Intrauterine pressure (IUP) is considered the “gold standard” to monitor uterine activity, though it requires membrane rupture and is highly invasive. Considering that uterine mechanical activity is a direct consequence of uterine myoelectrical activity, IUP signal can be non-invasively estimated from abdominal electrohysterogram (EHG) recordings. Previous works have reported EHG-based IUP estimates with linear parameters as root-mean-square or Teager energy. Due to non-linear nature of biological processes, the aim of this study was to test the performance of different non-linear EHG parameters to estimate IUP signal. Simultaneous EHG and IUP recordings were conducted in 17 women during labour. Teager energy (TE), Sample entropy (SampEn), Spectral entropy (SpEn), Lempel-Ziv (LZ), and Poincaré parameters: SD1, SD2, SDRR and SD1/SD2 were computed from the EHG. Different window lengths for computation and for a smoothing moving average filter were tested. Monovariable linear regression models were used to obtain IUP estimates. The best results were obtained with TE and SD1, both computed and filtered with windows of 5 s and 20 s, respectively. In the latter case, the RMSerror was 12.25 ± 4.03 mmHg, which points that non-linear EHG parameters can provide relevant information for non-invasive uterine activity monitoring.

Monica Albaladejo-Belmonte, Gema Prats-Boluda, Yiyao Ye-Lin, Carlos Benalcazar-Parra, Ángel Lopez, Alfredo Perales, Javier Garcia-Casado
Linear and Non-linear Analysis of EEG During Sleep Deprivation in Subjects with and Without Epilepsy

EEG has a central role in the diagnosis of epileptiform abnormalities helpful in diagnosing epilepsy. Since irregularities are random and sporadic events, easily activated in the initial phase of sleep but difficult to observe in a standard EEG examination, sleep deprivation is a frequent condition to be used. Thus, in this study the EEG monitoring of 44 subjects, 14 without epilepsy and 30 with epilepsy, afferent to the IRCCS Centro Neurolesi “Bonino Pulejo” of Messina were examined after sleep deprivation the day before performing the registration. EEGs were recorded according to the international setting system using nineteen channels. The normalized power spectral densities in delta (2–4 Hz), theta (4–8 Hz), alpha (8–13 Hz) and beta (13–30 Hz) band were computed and the non-linear parameters such as beta exponent, fractal dimension and zero crossing were considered. The differences between the sleep and awake were significant in almost all the channels in the beta band and in posterior areas for beta exponent, fractal dimension and zero crossing in normal subjects. In epileptic patients they were significant in all the channels in the delta band and for the non-linear parameters, and in several ones in theta and beta bands. Even if in posterior areas all the spectral and the non-linear parameters showed different values between epileptic and healthy subjects, no significant differences were found. The results suggest that analysis of spectral power as well as of complexity, obtained by non-linear parameters, could be used to identify differences between healthy and epileptic patients.

Silvia Marino, Giulia Silveri, Lilla Bonanno, Simona De Salvo, Emanuele Cartella, Aleksandar Miladinović, Miloš Ajčević, Agostino Accardo
Brain Oscillatory Activity and Neurological Deficit in Hyper-acute Ischemic Stroke: Correlation of EEG Changes with NIHSS

The accurate identification and prediction of cerebral infarct evolution and clinical outcome are of paramount importance in acute ischemic stroke management. Neuroimaging in acute stroke is mandatory to establish the feasibility of reperfusion therapy, but it is not practical to assess the continuous evolution of brain ischemia. EEG could be an applicable instrument to perform functional monitoring in the hyper-acute phase. EEG activity during ischemic stroke has been widely studied in sub-acute and post-acute phase of ischemic stroke. However, only few studies have focused on the early phase of brain ischemia. The aim of this study conducted at the stroke unit was to investigate stroke-related EEG changes during the earliest phase of ischemic stroke within 4.5 h from symptom onset and to correlate these data with neurological deficit in terms of NIHSS score. We studied 12 patients with ischemic stroke, who underwent EEG recordings within 4.5 h from symptom onset. The EEG signals acquisition was performed bedside without delaying reperfusion treatment, using @64 channels Wi-Fi Be Plus LTM amplifier and 19 channel 10–20 Ag/AgCl electrodes wireless prewired headset. The main finding of this study is a significant positive correlation between stroke-related EEG changes measured by DAR and DTABR parameters and the neurological deficit measured by NIHSS score, during the earliest phase of ischemic stroke. The results of this study highlight the importance of EEG as complementary tool in the assessment of stroke severity and its potential role in acute decision-making and monitoring.

Miloš Ajčević, Giovanni Furlanis, Lara Stragapede, Mariana Ridolfi, Paola Caruso, Marcello Naccarato, Agostino Accardo, Paolo Manganotti
Differences in Circadian Rhythms of Blood Pressure and Heart Rate Among Hypertensive and Normal Blood Pressure Subjects

The purpose of this study was to evaluate the differences in the circadian variations of blood pressure (BP) and heart rate (HR) among three subject groups (hypertensive, normal/high normal and optimal blood pressure). The ambulatory BP and HR of 385 subjects, without clinical evidence of hypertension-related complications, were acquired using a Holter Blood Pressure Monitor and their circadian patterns were analyzed. Systolic, Diastolic and Mean BP showed four different well-defined trends in specific time intervals of the day, similar for the three BP measures and among the three subject groups. Both BP and HR signals presented a decrease between 10:00 and 14:30 and from 19:00 to 2:00 and an increase from 5:00 to 10:00. Between 14:30 and 19:00, BP and HR presented an opposite relationship with decreasing HR and increasing BP, not yet reported in the literature. The behaviors of BP and HR were well approximated in each of the four periods by linear trends in all the three subject groups. On the contrary, in the period between 2:00 and 5:00 both BP and HR showed a quite constant trend. Results support the hypothesis of an independent vagal control during 24-h in respect of HR and BP mean levels, the latter depending on the specific subject group. Moreover, linear approximation in the identified four intervals could be used to quantify the circadian changes for all kind of subjects.

Silveri Giulia, Pascazio Lorenzo, Sabbadini Gastone, Guerra Monica, Accardo Agostino
Ectopic Beat Detection from Wrist Optical Signals for Sinus Rhythm and Atrial Fibrillation Subjects

Ectopic beats are abnormal cardiac beats originating from a location different than the sino-atrial node and therefore not being controlled by the autonomous nervous system. Thus, correct heart rate variability analysis inevitably requires accurate ectopic beat detection. Furthermore, an accurate ectopic beat detection is crucial to differentiate irregular cardiac rhythm due to different types of pathological arrhythmias from those caused by isolated ectopic beats. In this paper, we present an algorithm for ectopic beat detection based on wrist plethysmographic (PPG) signals. The proposed algorithm relies on analyzing the inter-beat patterns while considering the heart-rhythm condition; whether sinus rhythm (SR) or atrial fibrillation (AF). We monitor 29 patients recovering from surgery in the post-anesthesia care unit. During the recordings, 15 patients had SR and 14 patients had AF. The proposed ectopic beat detection algorithm achieves a sensitivity of $$93.08 \pm 3.83\%$$ and a specificity of $$97.80 \pm 2.12\%$$ .

Serj Haddad, Jarkko Harju, Adrian Tarniceriu, Tuomas Halkola, Jakub Parak, Ilkka Korhonen, Arvi Yli-Hankala, Antti Vehkaoja
Electrocardiographic Alternans: A New Approach

Alternans is an electrophysiological phenomenon consisting in a beat-to-beat variation of the morphology of an electrocardiographic (ECG) waveform. Literature has particularly studied T-wave alternans (TWA) because it has been widely recognized as a noninvasive and clinically useful index to predict occurrence of malignant ventricular arrhythmias and, eventually, sudden cardiac death. Historically, alternans of other segments of ECG, like P wave (PWA), or QRS complex (QRSA) gained less interest than TWA, but this is an incomplete vision of the action potential (AP). AP is influenced by electrical activity of all myocardial cells, so it is reasonable that all ECG waveforms could be affected by alternans phenomenon. ECG alternans (ECGA) can be intended as the prevalent nature of alternans. This study aimed to use the heart-rate adaptive match filter (AMF) method, previously applied for TWA applications, to detect ECGA. AMF effectiveness was tested on simulated alternating ECG (alternans-amplitude range: 10 µV–200 µV), characterized by single- and multiple-wave alternans (always of the same amplitude and morphology). AMF method proved to be specific, being able to recognize ECGA absence, and particularly sensitive to TWA. In general, in case of singular-wave alternans, AMF correctly identified the type of alternans and correctly determined its amplitude (mean error: 0%). When TWA was combined to PWA or QRSA, only TWA was identified with an overestimation of its amplitude (mean error: 23%). In conclusion, overall AMF proved its effectiveness and specificity in revealing and discriminating ECGA.

Ilaria Marcantoni, Dalila Calabrese, Giorgia Chiriatti, Roberta Melchionda, Benedetta Pambianco, Giulia Rafaiani, Eleonora Scardecchia, Agnese Sbrollini, Micaela Morettini, Laura Burattini
Co-activation of Knee Muscles in Female vs. Male Adults

During walking, knee joint mechanics is primarily regulated by thigh-muscle group, i.e. hamstrings and quadriceps femoris. Research purpose was to assess gender-related differences in concomitant recruitment of antagonist knee-joint muscles during ground walking. To this aim, Statistical gait analysis was performed on surface-electromyographic (sEMG) signals from vastus lateralis (VL) and medial hamstrings (MH) in 15 female (F-group) and 15 male (M-group) age-matched able-bodied young adults. sEMG signals from numerous strides (average value ± SD of 452 ± 102 strides for F-group and 440 ± 106 strides for M-group) were analyzed for each subject. Results showed that the same three VL/MH co-activations were found in the gait cycle, irrespective of gender: during early stance (ES), push-off (PO), and swing (SW) phase. No significant gender-related differences (p > 0.05) were observed in co-activity duration. Differently, an increase of occurrence frequency was observed in F-group for VL/MH co-activation during PO phase, with respect to M-group (21.9 ± 13.6% vs. 11.3 ± 8.6% of strides, p = 2.5 × 10−3). This increased occurrence of co-activations suggests a more complex muscular recruitment for knee-joint stabilization in females, in particular in PO phase when the control of balance is more awkward because of the final phase of single support. In conclusion, the present study indicates gender as a not negligible factor in evaluating knee-muscle co-activation during walking.

Francesco Di Nardo, Annachiara Strazza, Andrea Tigrini, Guido Mascia, Stefano Cardarelli, Alessandro Mengarelli, Federica Verdini, Sandro Fioretti
Automatic Segmentation of Bipolar EHGs’ Contractions Using Wavelet Transform

Until recently, many segmentation research trials on uterine EMG have been recorded for the sake of not only automatic detection of contractions but also curtailment of other events presents in the electrohysterogram (EHG). In this study, we use an online segmentation method, which has proven its efficiency and known by Dynamic Cumulative Sum (DCS). The method is first applied on real bipolar EHGs signals then on their details obtained by wavelet transform. The detected rupture instants are driven through an enhanced technique of faulty rupture instants elimination, dynamic selection of multichannel bipolar EHG signals and its details after wavelet transform and event tracking algorithm. Therefore, enhanced method sensitivity and “other events” rate of bipolar EHGs with and without wavelet are computed using Margin validation test in order to classify among events as contractions or not referring to contractions identified by expert. Indeed, enhanced technique of event tracking, proposed in this study, seems to be more efficient comparing to previous techniques. Further studies should be achieved for the sake of fully identifying the uterine contractions from other events and then decreasing the “other events” rate.

Amer Zaylaa, Ahmad Diab, Ziad Fawal, Mohamad Khalil, Catherine Marque
Methods for Removing of Line Noise Artifact from EEG Records with Minimization of Neural Information Loss

Line noise is artifact affecting a gamma band of the EEG. Conventional filters are used for removing the line noise artifact, but these filters, also remove physiological activity and could induce spurious oscillations. We tested two alternative methods for removing line noise artifact without removing a physiological activity, specifically Adaptive noise cancellation (ANC) method with Linear regression and modified Independent Component Analysis (ICA). We proposed a new protocol for statistical evaluation of line noise removal effectiveness. The line noise artifact was physically simulated and the protocol for statistical evaluation was verified. The ANC method shows good results according to statistical evaluation. The lowest residuum of simulated line noise artifact corresponds to 0.0005 $$\upmu $$ V $$^2$$ /Hz. Nevertheless, the ANC is very sensitive to the quality of the reference signal, which heavily depends on the reference electrode selection.

Jan Strobl, Marek Piorecky, Vlastimil Koudelka, Tomas Nagy, Vladimir Krajca
Pilot Study for Estimating Physical Fatigue Based on Heart Rate Variability and Reaction Time

The aim of this study was to evaluate how heart rate variability (HRV) and reaction time differ when measured in the morning and in the evening of one physically exhausting day in order to explore if these parameters would be sufficient for estimating physical fatigue accumulated during one day. Five different experiment days with fixed schedules were conducted which consisted of measurement set in the morning, physical exercise during the day and measurement set in the evening. The total average reaction time was lower in the morning (228 ± 18 ms) compared to the values measured in the evening (257 ± 22 ms). Both assessed HRV parameters SDNN and RMSSD showed a tendency to decrease during the day (total averages were respectively 61 ± 6 ms and 40 ± 6 ms in the morning vs 37 ± 4 ms and 24 ± 3 ms in the evening). This decrease was more prominent in the heart rate recovery phase compared to the resting heart rate. The results of this study give promising results for new methods for estimating daily physical fatigue and could be a basis for multiple future studies.

Ardo Allik, Kristjan Pilt, Moonika Viigimäe, Ivo Fridolin
Characterization of Eye Gaze and Pupil Diameter Measurements from Remote and Mobile Eye-Tracking Devices

Eye-tracking technology allows to capture real-time visual behavior information and to provide insights about cognitive processes and autonomic function, by measuring gaze position and pupillary response to delivered stimuli. Over the recent years, the development of easy-to-use devices led to a large increase in the use of eye-tracking in a broad spectrum of applications, e.g. clinical diagnostics and psychological research. Given the lack of extensive material to characterize the performance of different eye-trackers, especially latest generation devices, the present study aimed at comparing a screen-mounted eye-tracker (remote) and a pair of wearable eye-tracking glasses (mobile). Seventeen healthy subjects were asked to look at a moving target on a screen for 90 s, while point of regard (POR) and pupil diameter (PD) were recorded by the two devices with a sampling rate of 30 Hz. First, data were preprocessed to remove artifacts, then correlation coefficients (for both signals) and magnitude-squared coherence (for PD) were calculated to assess signals agreement in time and frequency domain. POR measurements from remote and mobile devices resulted highly comparable (ρ > 0.75). PD showed lower correlation and major dispersion (ρ > 0.50), besides a higher number of invalid samples from the mobile device with respect to the remote one. Results provided evidence that the two instruments do share the same content at the level of information generally used to characterize subjects behavioral and physiological reactions. Future analysis of additional features and devices with higher sampling frequencies will be planned to further support their clinical use.

Riccardo Lolatto, Giulia Rocco, Riccardo Mustoni, Chiara Maninetti, Riccardo Pastura, Andrea Pigazzini, Riccardo Barbieri
Efficacy of Time- and Frequency-Domain Heart Rate Variability Features in Stress Detection and Their Relation with Coping Strategies

Heart Rate Variability (HRV) is a reliable biomarker of the Autonomic Nervous System (ANS) activity and it is widely used to characterize the stress response induced by different laboratory stress tasks. Even if acute stress responses have been previously investigated, researchers are still wondering which HRV indices are more effective for stress assessment. In the present study, we extracted several time- and frequency-domain HRV parameters and investigated which ones are better able to discriminate between a stressful and a non-stressful condition. Moreover, we explored the possibility of a linear correlation between such parameters and certain coping strategies, during three laboratory stress tasks (Montreal Imaging Stress Task, Stroop Color-Word, Speech). The effect size computed for each considered HRV parameter indicate that the average RR interval and the normalized power in low frequency (LF) band were the most effective parameters for the detection of mental stress. As regards the second hypothesis, normalized power in high frequency (HF) band during Speech was found significantly and negatively correlated with specific subscales of the administered questionnaires (CERQ-S and COPE-NVI), suggesting a possible association between the higher use of social support (SS) and other blame (OB) coping strategies and stronger autonomic responses during the Speech task.

Pierluigi Reali, Agostino Brugnera, Angelo Compare, Anna Maria Bianchi
Influence of Physical Models of Electrodes on Rat’s Head Forward Modelling

Reliable inverse imaging of source currents in rat’s brain requires accurate models of fields and interfaces. Accuracy of field models can be influenced by a proper representation of electrodes used for sensing potentials. In the paper, we study the effect of realistic modelling of electrodes on the distribution of electric field intensity in the head. Fields are evaluated for three commonly used configurations of electrodes using a three-dimensional finite element model of rat’s head comprising the brain, the cerebrospinal fluid, and the skull. Simulations show a lower amplitude of electric field intensity if realistic models of electrodes are considered. Moreover, electric field intensity near electrode models is shown to be slightly influenced. Therefore, realistic modelling of electrodes can improve the accuracy of an inverse imaging, and consequently make the localization of current sources in rat’s brain more accurate.

David Kuratko, Jaroslav Lacik, Zbynek Raida, Daniel K. Wójcik, Vlastimil Koudelka
Improvement of Sleep Spindle Detection by Aggregation Techniques

The study focuses on automatic sleep spindle detection. Plenty of methods have been proposed in previous decades. However, there is still space for improvement. In this study, we investigate aggregation methods such as voting to achieve better results. We employ an unweighted model and two weighted voting models in which assigned weights represent reliability of automatic detectors. First weighted model utilizes supervised approach based on logistic regression. The second one applies unsupervised generative Bayesian model often used in crowdsourcing. Using the expectation maximization algorithm, we uncover hidden true labels and weighs of detectors. We test methods on the real world datasets. The aggregation method overcome single detectors on 10% on average in terms of F1. Moreover, a probabilistic explanation of weights could be used in applications for visual analysis.

Elizaveta Saifutdinova, Daniela Dudysova, Vaclav Gerla, Lenka Lhotska
Preprocessing Pipeline for fNIRS Data in Children

Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging technique, largely used in paediatric research. However, there is not a standardized and widely accepted protocol for fNIRS data processing with potential effects on the reliability and replicability of the obtained results. The present study is within this framework aiming at the identification of an adequate pre-processing pipeline to be used for the analysis of children fNIRS datasets. The performance of five different motion correction techniques, based on the principal component analysis and on the wavelet filtering, was evaluated by analyzing fNIRS data recorded in 22 typically developing children (mean age 11.4 years). The results showed that the wavelet analysis combined with a moving average filter achieved the best performance, suggesting that this technique might become a gold-standard approach for motion artifacts correction in fNIRS children’s datasets.

Caterina Piazza, Andrea Bacchetta, Alessandro Crippa, Maddalena Mauri, Silvia Grazioli, Gianluigi Reni, Maria Nobile, Anna Maria Bianchi
Wavelet Analysis-Based Reconstruction for sEMG Signal Denoising

Surface electromyography (sEMG) recordings provide a safe, easy, and non-invasive method, allowing objective quantification of the electric activity of muscles. Analysis of sEMG plays an important diagnostic role in assessing muscle disorders. Typically, sEMG is a non-stationary signal contaminated by various noises or artifacts that originate at the skin-electrode interface, in the electronics, and in external sources. Thus, appropriate filtering procedures have to be applied to make sEMG clinically usable, in order to extract the main sEMG features. In the recent literatures, among the best performing denoising methods, Wavelet transformation (WT) denoising has been proposed. In particular, aim of this study is to propose a new denoising method based on WT multi-level decomposition analysis. To this aim, Daubechies mother wavelet (4th order, 9 levels of decomposition) was applied to 5 real sEMG tracings. Tibialis anterior (TA) and gastrocnemius lateralis (GL) signals are considered. This method focusses on the choice of a new thresholding rule for sEMG reconstruction and denoising. Performances of this method are computed against soft-thresholding denoising technique (ST) in terms of Root Mean Square Error (RMSE). After application of WT multi-level denoising technique, signal-to-noise ratio (SNR) increased significantly (TA: 14.5 ± 6.9 vs. 19.5 ± 7.1; GL: 14.0 ± 5.4 vs. 18.7 ± 6.3). Moreover, WT multi-level denoising technique showed a lower dispersion than ST (RMSE for TA: 0.8 vs. 1.2; RMSE for GL: 0.9 vs. 1.1.), introduced no sEMG signal delay. Thus, this method is a novel and efficient tool for sEMG denoising, that could be used to make easier the detection of sEMG activation onset-offset.

Annachiara Strazza, Federica Verdini, Alessandro Mengarelli, Stefano Cardarelli, Andrea Tigrini, Sandro Fioretti, Francesco Di Nardo
An Information-Theoretical Method for Emotion Classification

Identifying the emotion that someone is feeling will allow to improve the experience of the person interaction with environments, devices, and contents. Our body responds to events around us, by emotional responses, reflected in cognitive, behavioral and physiological dimensions. In the present work, we target the electrocardiogram (ECG) response as a mean to express emotions. Its processing is performed using information-theoretical measures, allowing true exploratory data mining. Participants recruited for the experiment watched three video sets in three different days, with a different emotion being induced in each day: fear, happiness, and neutral condition. The method is divided in: (1) conversion of the real-valued ECG record into a symbolic time-series; (2) relative compression of the symbolic representation of the ECG, using the symbolic ECG records stored in the database as a reference; (3) identification of the ECG record class, using a 1-NN (nearest neighbor) classifier. An accuracy of 90% was obtained. A posteriori analysis of the false negative results indicated that there was a relation between the relative dissimilarity measure and the self-reported emotions.

Susana Brás, João M. Carvalho, Filipa Barros, Claúdia Figueiredo, Sandra C. Soares, Armando J. Pinho
Potential Biomechanical Overload on Skeletal Muscle Structures in Students During Walk with Backpack

Although a very large number of students in the world use uncomfortable and heavy backpacks, their negative musculoskeletal effects on gait and posture are still not well investigated. Aim of the paper has been the study of differences affecting the kinematic gait parameters during free walk and walk with backpack to evaluate their potential influence on skeletal-muscle disorders. Gait recordings in both conditions on 50 healthy volunteers participating students have been performed by a G-WALK inertial system calculating the main kinematic parameters namely Propulsion index and Initial Double Support, Stance and Swing Phases. ANOVA results between both walking conditions showed that all gait cycle studied values are significantly negatively affected by walking with backpack supposing a potential biomechanical overload on skeletal muscle structures in students exposed to these prolonged conditions.

Giovanni D’Addio, Leandro Donisi, Luca Mercogliano, Giuseppe Cesarelli, Paolo Bifulco, Mario Cesarelli
Accurate Calculation of Heart Period and Pulse Wave Transit Time

The first step in processing of physiological signals is usually the improvement of the signal-to-noise ratio. However, the applicable filtering methods not only suppress noise but also distort the signal. As a result, parameters derived from filtered physiological signals are also distorted. The paper evaluates filtering methods applied to electrocardiographic (ECG) and photoplethysmographic (PPG) signals in the process of calculating heart period (HP) and pulse wave transit time (PWTT). Band-pass filtering applied in the course of QRS detection causes a shift in R-peak location depending on the QRS waveform. Following QRS detection accurate R-peak localization is suggested using re-filtering of the original signal. The HP was calculated as the time interval between two adjacent R-peaks and also as the time interval between two adjacent local minima in PPG (PPGmin). Both real and simulated data were used for the evaluation. In the latter case the fiducial points of the signals (ECG R-peak and PPGmin locations) were exactly known providing reference positions for the evaluation of filter distortion. Inappropriate filtering shifts the fiducial points. This causes a significant error while calculating parameters by subtraction, like RR interval of heart cycles (tRR, difference of two consecutive R-peak locations) or PWTT (difference of PPGmin and the associated R-peak location). Even more accurate R-peak and PPGmin localization is needed when variation in tRR and PWTT is analyzed. These variations characterize the actual stress level of a person before or during blood pressure measurement more accurately than average heart rate displayed by commercially available blood pressure monitors. Based on accurate R-peak and PPGmin localization it has been confirmed that for the same heart cycle the time interval is different between two R-peaks and two PPGmins.

Péter Nagy, Ákos Jobbágy
Long-Term Stability of EEG Spectral Asymmetry Index – Preliminary Study

The purpose of this preliminary study was to assess the long-term stability of the electroencephalography (EEG) based spectral asymmetry index (SASI) previously proposed as an objective measure for evaluation of depression. Due to high variability between individuals, a long-term study is required to examine how SASI changes over longer period of time for an individual subject. One healthy adult was surveyed over the period of 15 months. The eyes closed resting state EEG was analyzed over 15 sessions, on average once a month. The SASI was calculated from EEG power spectrum density as a relative difference in powers of higher and lower band of the EEG spectrum maximum. The long-term stability of SASI for an individual was compared to inter-individual variability of the measure within the group of healthy subjects in our previous study. The results demonstrated that the individual long-term variability of SASI is less than inter-individual variability within a group, supporting the possibility of application of SASI for evaluation of depression symptoms for an individual. In future, the investigations should be extended on larger number of subjects to study the individual properties of SASI.

Tuuli Uudeberg, Laura Päeske, Toomas Põld, Jaanus Lass, Hiie Hinrikus, Maie Bachmann

Regular Sessions: Biomedical Imaging and Image Processing

Flow Convergence Area Estimation on In Vitro Color Flow Doppler Images Using Deep Learning

We present an automatic method to estimate flow rate through the orifice in in-vitro 2D color-flow Doppler echocardiographic images. Flow rate properties are important for the assessment of pathologies like mitral regurgitation. We expect this method to be transferable to in-vivo patient data. The method consists of two main parts: (a) detecting a bounding box which encloses aliasing contours and its surroundings (namely a region representative of flow convergence area), (b) application of Convolutional Neural Networks for regression to estimate the flow convergence area. Best result achieved is the 5% mean error for validation data which is from other experiments that were used for training. Given the small number of training data, this method shows promising results.

Grigorios-Aris Cheimariotis, Kostas Haris, Jeesoo Lee, Brent E. White, Aggelos K. Katsaggelos, James D. Thomas, Nikolaos Maglaveras
Automated Design of Efficient Supports in FDM 3D Printing of Anatomical Phantoms

Recent improvements in image segmentation techniques enabled the (semi)automatic extraction of biostructures surfaces from 3D medical imaging data. The diffusion of 3D printing technologies promoted their introduction in the medical field, giving rise to several applications, such as the development of 3D-printed anatomical imaging phantoms. These devices provide controlled experimental environments for the improvement of medical imaging techniques, as they mimic the morphological and physiological features of different body parts. However, to obtain a 3D printable model from medical imaging data, different post-processing steps are needed, which require a considerable effort. Supports generation is often a critical task, as it requires to find the minimum amount of support structures necessary to hold a part in place during the printing process. This is particularly difficult for complex anthropomorphic models, for which a high printing level of detail, along with a reasonable number of internal supports, is usually needed. In this paper, an automatic method for the design of efficient support structures is proposed, which is suitable for 3D printing of complex anatomical phantoms, even with non-professional FDM 3D printers. A custom design software was developed, which places paraboloid-shaped shells to support all and only the critical points of the 3D model. This provided different advantages over support generation by means of common slicing software, allowing a reduction of material waste and printing times, along with an easier and faster dissolution of soluble supports for the clean-up of phantoms empty volumes.

Maria Agnese Pirozzi, Emilio Andreozzi, Mario Magliulo, Paolo Gargiulo, Mario Cesarelli, Bruno Alfano
Diffusion Weighted Magnetic Resonance Imaging Texture Biomarkers for Breast Cancer Diagnosis

Quantification of breast lesion heterogeneity by means of MRI texture contributes in differentiating benign from malignant breast lesions. This study investigates the diagnostic performance of 1st and 2nd order Texture Analysis descriptors on Apparent Diffusion Coefficient (ADC) lesion maps. 78 histologically verified breast lesions (40 benign, 38 malignant) of 67 patients undergoing DW-MRI at 3.0 T, were analyzed. ADC maps were generated for a slice representative of lesion largest diameter. A two-step segmentation approach was applied on high b-value diffusion image, based on Fuzzy C-Means (FCM) clustering and edge-based contouring, for defining the lesion region contour. Lesion contour was transferred to ADC map and subjected to texture analysis by means of twelve first-order and eleven second-order texture features. Logistic Regression Classifier was employed to assess the diagnostic ability of individual features and feature combinations. Diagnostic performance was evaluated by means of the area under Receiver Operating Characteristic curve (Az). The highest classification performance (Az = 0.965 ± 0.024) was achieved by the combined feature subset 25th Percentile (1storder) and Entropy (2ndorder), suggesting the diagnostic significance of accurately quantifying lesion heterogeneity by texture-based feature combinations on ADC maps. Combined 1st and 2nd order texture biomarkers provide accurate spatial information of lesion ADC heterogeneity and holds potential in differentiating benign from malignant breast lesion status.

Marialena I. Tsarouchi, Georgios F. Vlachopoulos, Anna N. Karahaliou, Lena I. Costaridou
Modeling Functional Processes of Brain Tissue: An fMRI Study on Patients with Un-Medicated Late-Onset Restless Leg Syndrome

In the current study we focus on the modeling of functional processes of brain tissue using functional Magnetic Resonance Imaging (fMRI) data. A brain connectivity analysis of Restless Legs Syndrome (RLS) is presented. Thirteen un-medicated patients with late-onset RLS and six healthy subjects are studied using structural and functional brain images. We compare functional connectivity analysis methods, according to their dependency on models or data, as well as to model effective connectivity. An Independent Component Analysis (ICA) method is implemented and all spontaneously activated areas in resting-state condition in both patients and healthy subjects are recorded and be compared with previous studies. Functional connectivity correlation matrices of both RLS and control subjects are extracted and these functional connectivity measures were compared using a seed-based analysis method. We model the brain tissue, based on the influence that one region exerts over another, using a spectral Dynamic Causal Model (DCM) analysis, which has not yet been implemented for RLS data. Finally, a Bayesian Model Selection is chosen in order to compare the winning model that effectively describes the data. The benefits of each methodology are presented.

Amalia K. Ntemou, Evanthia E. Tripoliti, Persefoni N. Margariti, Maria I. Argyropoulou, Dimitrios I. Fotiadis
Shift-Compensated Volumetric Interpolation of Tomographic Sequences for Accurate 3D Reconstruction

In patients affected by craniosynostosis, i.e. a congenital cranial defect, diagnostic evaluation for a prompt surgical treatment is performed using low-dose three-dimensional computer tomography (CT), characterized by a poor spatial resolution (in terms of slice thickness). The limited number of CT images reduces the accuracy of the 3D reconstruction of the skull and leads to a coarser segmentation and modelling. In this paper, Motion Compensated Frame Interpolation (MCFI) techniques are applied for an effective axial interpolation of tomographic images sequences, with the main objective of obtaining a refined 3D reconstruction. The performance of the proposed method was assessed by using high-resolution CT sequences. After downsampling along the axial direction, the missing slices were recovered by using the proposed algorithm, to obtain an estimate of the original sequence. The experimental results show that the 3D models obtained from the downsampled/interpolated sequence are very close to those obtained from the original one thus providing a high-quality 3D skull reconstruction.

Chiara Santarelli, Francesca Uccheddu, Fabrizio Argenti, Luciano Alparone, Monica Carfagni, Lapo Governi
Calculating Texture Features from Mammograms and Evaluating Their Performance in Classifying Clusters of Microcalcifications

In this work, 2432 texture features were calculated from microcalcification clusters presented on 190 images from the Digital Database for Screening Mammography. Mutual information technique was used to rank texture features. Then, an incremental procedure adds top ranked features to the Fisher discriminant analysis to determine the best set of texture features in classifying benign or malignant microcalcification clusters. The result was achieved using 13 texture features (AUC.632+ = 0.945 ± 0.019). However, to assure a consistent statistical analysis, at least 30 sample images for each feature added was assumed. The best performance was achieved by a set with 5 texture features (AUC.632+ = 0.884 ± 0.025), which is comparable to the ones presented in literature.

Marcelo A. Duarte, Wagner C. A. Pereira, André Victor Alvarenga
LNDetector: A Flexible Gaze Characterisation Collaborative Platform for Pulmonary Nodule Screening

Lung cancer is the deadliest type of cancer worldwide and late detection is one of the major factors for the low survival rate of patients. Low dose computed tomography has been suggested as a potential early screening tool but manual screening is costly, time-consuming and prone to interobserver variability. This has fueled the development of automatic methods for the detection, segmentation and characterisation of pulmonary nodules but its application to the clinical routine is challenging. In this study, a platform for the development, deployment and testing of pulmonary nodule computer-aided strategies is presented: LNDetector. LNDetector integrates image exploration and nodule annotation tools as well as advanced nodule detection, segmentation and classification methods and gaze characterisation. Different processing modules can easily be implemented or replaced to test their efficiency in clinical environments and the use of gaze analysis allows for the development of collaborative strategies. The potential use of this platform is shown through a combination of visual search, gaze characterisation and automatic nodule detection tools for an efficient and collaborative computer-aided strategy for pulmonary nodule screening.

João Pedrosa, Guilherme Aresta, João Rebelo, Eduardo Negrão, Isabel Ramos, António Cunha, Aurélio Campilho
Physical Breast Phantom Dedicated for Mammography Studies

This paper presents a simple methodology for creation of anthropomorphic physical breast phantoms dedicated for x-ray breast imaging. The use of physical phantoms in diagnostic radiology is a well-established approach for patient dose estimation, quality control of developed diagnostic systems and development of new breast imaging techniques. An overview of the design process used for the creation of the presented phantom and evaluation approach is given. A combination of materials (Clear resin, Gray resin and PLA) and methods (fused deposition modelling and stereolithography) are used for the manufacturing of the breast phantom. The phantom was subjected to an evaluation aiming at its suitability for studies with digital mammography technique. In particular, the created phantom was evaluated by using two sets of phantom images, taken using 22 kVp and 28 kVp, which are compared with real mammograms. The comparison of the images is based on extracted statistical parameters – namely skewness, kurtosis, fractals, PSA, GLCMContrast and GLCMEnergy.

Firgan Feradov, Stoyko Marinov, Kristina Bliznakova
Segmentation of Pulmonary Nodules in CT Images Using the Sliding Band Filter

This paper proposes a conventional approach for pulmonary nodule segmentation, that uses the Sliding Band Filter to estimate the center of the nodule, and consequently the filter’s support points, matching the initial border coordinates. This preliminary segmentation is then refined to try to include mainly the nodular area, and no other regions (e.g. vessels and pleural wall). The algorithm was tested on 2653 nodules from the LIDC database and achieved a Dice score of 0.663, yielding similar results to the ground truth reference, and thus being a promising tool to promote early lung cancer screening and improve nodule characterization.

Joana Rocha, António Cunha, Ana Maria Mendonça
Method for Finding the Limits of Blood Vessel Landmarks in Eye Fundus Images Based on Distances in Graphs: Preliminary Results

The paper proposes a method to process blood vessel landmarks in eye fundus images. The proposed method is based on finding a shortest cycle in a weighted graph, corresponding to a set of possible tracked blood vessel slices and paths between them. In turn, the paths are found by finding a shortest path in a weighted graph, taking gradients into account. The method has been tried out with DRIVE and IOSTAR databases.

Martynas Patašius, Jūratė Šimkienė, Daivaras Sokas, Andrius Pranskūnas
Anthropomorphic Physical Breast Phantom Based on Patient Breast CT Data: Preliminary Results

We propose a novel approach for the fabrication of realistic in shape, size as well as in x-ray absorption properties 3D physical breast models. The approach eliminates the need for segmentation of breast tissues by directly mimicking the radiodensity of each voxel in a computed tomography (CT) image. This is done through a recently published study that enables the use of variable filament extrusion rate when creating 3D phantoms with fused deposition modeling printers. We use the CT image of a real patient’s breast to produce a model through the suggested approach. We then validate the fabricated physical breast phantom by obtaining a CT scan of it and comparing the latter to the original CT image of the source patient’s breast.

Sivo Daskalov, Nikiforos Okkalidis, John M. Boone, Stoyko Marinov, Zhivko Bliznakov, Giovanni Mettivier, Hilde Bosmans, Paolo Russo, Kristina Bliznakova
Microcalcification Cluster SDNR in Synthesized and 2D Mammography

This study aims at quantitative assessment of image quality of microcalcification clusters in synthesized-2-Dimensional (s2D) mammography (i.e. generated from digital breast tomosynthesis-DBT) as compared to 2-Dimensional (2D) mammography, in terms of the signal-difference-to-noise ratio (SDNR). A mammographic phantom (Leeds, TORMAM) simulating microcalcification (MC) clusters located in varying parenchyma density patterns (fatty, glandular and dense) was imaged in 2D and DBT, considering two distinct scattering conditions (phantom thickness 30 mm and 50 mm). An accurate MC cluster segmentation method was used to delineate individual MC cluster particles and define overall cluster SDNR in s2D and 2D images. Pairwise comparisons of each cluster SDNR between 2D and s2D demonstrated no statistically significant difference across the three parenchyma density patterns (p > 0.05, Mann Whitney U test), for both scattering conditions. s2D demonstrated constant performance in terms of cluster SDNR with increasing surrounding parenchyma density, while in case of 2D, cluster SDNR is decreasing with substantial parenchyma density increase.

Andreas Petropoulos, Spyros Skiadopoulos, Anna Karahaliou, Georgios Vlachopoulos, Gerasimos Messaris, Lena Costaridou
Enhancing CT 3D Images by Independent Component Analysis of Projection Images

Computed tomography (CT) is an imaging modality producing 3D images from sets of 2D X-ray images taken around the object. The images are noisy by nature, and segmentation of the 3D images is tedious. Also, detection of low contrast objects may be difficult, if not impossible. Here, we propose an independent component analysis (ICA) based method to process sets of 2D projection images prior to 3D reconstruction to remove noise, and to enhance objects for detection and segmentation. In this paper, a proof-of-concept is provided: the proposed method was able to separate noise and image components, as well as to make visible objects that were not observable in 3D images without processing. We demonstrate our method in object separation with 2D slice image processing simulations, and by enhancing a 3D image of a polymer sample taken with Xradia MicroXCT-400. The method is applicable in any CT tomography for which a number of project image sets with different contrasts can be taken, e.g., in multispectral fashion.

Markus Hannula, Jari A. K. Hyttinen, Jarno M. A. Tanskanen
Potentials of OCT in Monitoring Ocular Hemodynamics of Patients with Primary Open Angle Glaucoma

Approaches to using optical coherence tomography (OCT) for ocular hemodynamics monitoring in patients with primary open angle glaucoma (POAG) are presented. The existing OCT systems used in ophthalmology are described. The different OCT systems are compared in accuracy and resolution characteristics, as these are important for clinical studies. The results of ocular hemodynamics monitoring of POAG patients using Spectralis OCT2 with an OCT Angiography module are given.

E. N. Iomdina, D. D. Khoziev, A. A. Kiseleva, P. V. Luzhnov, O. A. Kiseleva, D. M. Shamaev
Automatic Segmentation of Bone and Muscle Structures in CT Volumes Using Convex Relaxation and Fine-Tuning

Segmentation of muscle and bone structures in CT volumes is a complicated task for physicians and surgeons due to the similarity of Hounsfield values with those within surrounding organs. The lack of efficient and automatic tools for this task makes that most physicians and surgeons prefer to segment bone and/or muscle tissue manually rather than working with specific-focused-tissue segmentation tools. In this work, an automatic tool for the joint segmentation of muscle and bone structures is described. The tool implements the segmentation in two stages. In the first stage an energy minimization approach via convex relaxation is performed. In the second step, a more accurate muscle segmentation is achieved using probability maps. A public database of CT volumes have been used to test the algorithm and different metrics, including Dice, Jaccard, sensitivity and specificity, have been calculated to evaluate the algorithm performance. The results provided outperform recent methods using the same public database.

José-Antonio Pérez-Carrasco, Carmen Serrano, Begoña Acha
A Comparison of Denoising Algorithms for Effective Edge Detection in X-Ray Fluoroscopy

X-ray fluoroscopy provides various diagnosis and is widely used in interventional radiology. However, the low-dose involved in fluoroscopy generates an intense Poisson-distributed quantum noise. Object recognition and tracking help in many fluoroscopic applications. Edge-detection is essential, but common derivative operators require noise suppression to provide reliable results. Moreover, homoscedasticity of noise is generally assumed, but is not the case of fluoroscopic images. However, the Anscombe transform can stabilize the quantum noise variance. This study presents a comparison of two denoising algorithms to evaluate their performance in edge-detection for real fluoroscopic sequences. VBM4D is one of best video-processing method for Additive White Gaussian Noise (AWGN), while Noise Variance Conditioned Average (NVCA) is a recent, real-time, algorithm specifically tailored for fluoroscopy. Some real fluoroscopic sequences screening the motion of lumbar spine were processed. Noise parameters were estimated using image sequences of a static scene: the relationship between the luminance and the noise variance was obtained. Generalised Anscombe transform and its inverse were applied to use the VBM4D algorithm. Edge-detection was performed by means of the Sobel operator. The Anscombe transform resulted able to stabilise the noise variance and consequently allow the use of algorithms designed for AWGN. The results show that both approaches provide effective identification of object contours (i.e. vertebral bodies). Despite of its simplicity the NVCA algorithm shows better performances than VBM4D on delineation of boundaries of examined spine fluoroscopic scenes. Furthermore, the NVCA algorithm can be realized in hardware and can offer real-time fluoroscopic processing.

Emilio Andreozzi, Maria Agnese Pirozzi, Antonio Sarno, Daniele Esposito, Mario Cesarelli, Paolo Bifulco
Stereophotogrammetric Basic Framework for Postural Assessment

Deviations of the vertebral column have become increasingly common. The follow-up of a spine deviation is performed every 2 or 3 months by routine visits and X-ray examinations. X-ray examination produces ionising radiation, which, over time, can cause cancer and is accumulative: this risk increases with the number of exposures. This work describes the conception and initial development of a simple and portable system based on stereoscopy capable of performing a topographic analysis of the patient’s back, as well as 3D reconstruction, in order to subsequently allow postural assessment. The stereoscopic system was built and calibrated, then the images underwent a process of rectification and point matching to create a depth map, so that preliminary results point to adequate 3D reconstruction of the back of a plastic phantom. The advantage of this method is the minimization of parallax errors, low cost, easy implementation and portability.

Alice Fontes, Mauricio Cagy
Dermoscopic Image Segmentation: A Comparison of Methodologies

An accurate segmentation of pigmented lesions may improve classification results of Computer Aided Diagnosis (CAD) tools. Thus, finding a reliable segmentation methodology becomes crucial. During the past few years, many segmentation methodologies of dermoscopic images have been proposed. In this paper, a comparison between three methodologies is presented: semantic segmentation with SegNet, histogram-based segmentation via convex optimization and segmentation based on a Fully Convolutional Network (FCN). As a result of evaluating the segmentation results for 600 dermoscopic images from the Test set of ISIC-2017 database, the semantic segmentation provides a 90.12% of accuracy, followed by segmentation using histograms and Fully Convolutional Network, with 86,47% and 81,70% of accuracy, respectively.

Paulina Vélez Núñez, Carmen Serrano, Begoña Acha, José Antonio Pérez-Carrasco
Quantitative Analysis of Brain 18F-fluordesoxyglucose and Early-Phase 18F-florbetapir Positron Emission Tomography

Positron emission tomography (PET) with amyloid binding tracers and 18F-fluordesoxyglucose (FDG) is used in the diagnosis of Alzheimer’s disease and other dementias. The dual-phase amyloid PET protocol acquires an early perfusion image immediately after the radiotracer injection that resembles FDG images. We studied the correlation of early phase 18F-florbetapir (eFBP) and FDG PET images by the quantitation of regional radiotracer uptake. Subjects with a neurodegenerative disease and eFBP and FDG PET scans were retrospectively selected. The PET images were normalized to the Montreal Neurological Institute template using Statistical Parametric Mapping version 12. Regional standardized uptake value ratios (SUVR) were extracted with the cerebellum serving as the reference region. Pearson’s correlation tests were performed to evaluate the regional correlation between eFBP and FDG PET images based on the quantitative measurements. Overall high correlation coefficients of up to r = 0.95 were observed in cortical regions that are in line with other studies or exceed previous results.

Alexander P. Seiffert, Adolfo Gómez-Grande, Patricia Sánchez-González, Walid Dghoughi, Alberto Villarejo-Galende, Héctor Bueno, Enrique J. Gómez

Regular Sessions: Bioinstrumentation, Biosenso and Bio-micro/nano Technologies

Cardiac Pacemaker Exposed to Electroporation Pulses – An Ex Vivo Study

Electrochemotherapy (ECT) consists of administration of a chemotherapeutic drug followed by local application of high-voltage electroporation (EP) pulses and is a well-proven method for treatment of various types of tumors. However, as a precaution ECT of deep-seated internal tumors is contraindicated in patients with implanted cardiac pacemakers. To address this limitation a functioning pacemaker programmed in asynchronous D00 mode with the ventricular and atrial leads connected was exposed to EP pulses in a conductive medium mimicking the conditions encountered in clinical setting. EP pulses were generated by the Cliniporator Vitae device and delivered via electrodes for clinical ECT. Application of EP pulses was synchronized with different phases of the pacing cycle, including the potentially most vulnerable phase, the pacing pulse itself. The effects on pacemaker operation were investigated by recording the voltages between the cathode and the anode of the bipolar ventricular lead. EP pulses induced large artefacts far exceeding the amplitude of the pacing pulse itself. This was followed by transiently disrupted shape of the remaining portion of the pacing pulse. However, all effects were strictly limited to this single pacing pulse. No effects at all were observed in subsequent pacing pulses. The pacemaker and its function remained completely immune to electroporation pulses delivered outside the pacing pulse itself.

Tomaz Jarm, Tadej Krmac, Damijan Miklavcic, Ratko Magjarevic
Smart Vest for Respiratory and Physical Activity Monitoring in COPD Patients

Rehabilitation treatment of patients with chronic obstructive pulmonary disease (COPD) based on muscle training provides clear benefits in health and quality of life. Its positive effects are usually diminished once the training is over, and the effectiveness of home maintenance programs is limited due to the lack of motivation and adherence of patients, as well as by the reduction of clinical supervision. In this work, the preliminary design and first results of a smart vest for the motivation and evaluation of patients with COPD in home maintenance programs are presented. The physical activity (metabolic expenditure and number of repetitions of the exercises) and the respiratory pattern of the patient are the parameters monitored through the built-in sensors in the smart vest. The processing of the information is distributed between the smart vest and a smartphone, which also acts as user interface and link for the management and remote data storage. Bidirectional wireless communications between the smart vest and the smartphone are based on a waveform preprocessing protocol over the standard Bluetooth 4.1, whose purpose is to reduce energy consumption. A preliminary study with healthy volunteers in a controlled environment has shown the feasibility of the proposed solution (100% accuracy in the number of exercise repetitions, close correspondence between the monitoring tests of metabolic expenditure and breathing and the pre-established patterns).

David Naranjo-Hernández, Javier Reina-Tosina, Laura M. Roa, Gerardo Barbarov-Rostán, Alejandro Talaminos-Barroso, Pilar Cejudo-Ramos, Eduardo Márquez-Martín, Francisco Ortega-Ruiz
A Prototype of Intelligent Portable Oxygen Concentrator for Patients with COPD Under Oxygen Therapy

Traditional long-term oxygen therapy devices used in daily-life activities need to be manually adjusted depending on the level of intensity of the activity done. In this study, a system to automatically control a commercial POC devices is proposed with the aim of improving the used and efficacy of the therapy in patients with chronic obstructive pulmonary disease.A patient unit and a control unit were designed and implemented to control a Inogen G2 oxygen concentrator. The patient unit included an inertial measurement unit to classify in real-time the physical activity performed by the user. The control unit receive via Bluetooth (BLE) the degree of activity and adjusted the oxygen concentrator automatically following the setting provided by clinician.Data from a group of six volunteers were used to train and validated the system using different machine learning algorithms. A logistic regression model (LR) achieved the best accuracy (93.69%) in classifying the physical activity and consequently adjusting correctly the O2 dose.The developed system may improve oxygenation through automatic adaptation to lifestyle, could contribute to improve the patient’s therapeutic compliance and may promote physical activity.

Alejandro Lara-Doña, Daniel Sanchez-Morillo, María Pérez-Morales, Miguel Ángel Fernandez-Granero, Antonio Leon-Jimenez
Optical Metrology of Novel Optically Stimulated Semiconductor Gas Sensor

Modern semiconductor sensors which are used for gas detection are based on an MOS (metal–oxide–semiconductor) structure, but the usage of such sensors is limited due to its working principles. Over the last decade, MOS sensors have been combined with a UV-radiation source, but the sensing capabilities of a low gas concentration for those sensors are limited. In previous studies, the possibility to use optically stimulated semiconductors to sense acetone gas at lower concentrations was demonstrated [1]. Another study demonstrated the possibility to sense and determine various solvent gases, where acetone vapour showed the highest signal increase [2].In this paper, the influence of optical stimulation parameters (irradiance and wavelength) on the output current change and saturation time for an optically stimulated semiconductor gas sensor was studied. It was shown that the maximal deviation of the output current change is achieved at lower tested irradiance Φ = 1.6 mW·cm−2, and the saturation time decreases with the increase of irradiance. The use of a 367.5 nm LED demonstrated a higher output current change. The saturation time decreased with the increase of the wavelength.

Yuri Dekhtyar, Maksims Komars, Maksims Sneiders

Regular Sessions: Bioinformatics, Computational Biology and Systems Biology

Dose–Response Curve: Temporal Dynamics of Respiratory Mechanics in Mice

A very usual practice in literature associated to respiratory mechanics is the challenge of the system, in order to evidence potential differences. Therefore, dose-response curves allow the assessment of respiratory mechanics. Despite the fact that the study of the dose-response curve using peak responses is important and desirable, an evaluation of the temporal dynamics of each dose in a dose-response curve could bring new perspectives about respiratory mechanics. Therefore, the present study aims to understand the kinetics of Newtonian resistance (Rn) of the respiratory system due to transient bronchoconstriction induced by administration of methacholine (MCh) in different concentration. Nine-week female Balb/c mice were studied, nine of the control group and ten of the lung inflammation group (OVA). The doses of MCh were 30, 100 and 300 μg/kg. The constants modelling the single exponential function were compared among of groups and doses. As a result, the increasing values of fitting constants are in direct relation to increasing values of MCh infusion. Additionally, the decay of the curve did follow the increase of the peak value for OVA group. However, it was possible to realize statistically that the temporal dynamics response of bronchoconstriction differs from control to OVA groups only at MCh dose of 100 μg/kg.

Otavio Henrique F. Ledesma, Renato L. Vitorasso, Maria Aparecida de Oliveira, Henrique Takachi Moriya
Influence of Astrocytic Gap Junction Coupling on in Silico Neuronal Network Activity

Astrocytes cover a plethora of roles supporting neurons in their maturation and regulating the concentrations of several ions and neurotransmitters. Moreover, astrocytes dysfunctions are, nowadays, suspected to have important implication in several brain diseases, as for example in epilepsy and Alzheimer’s disease. The astrocytes themselves are forming a network mediated by gap junctions. A loss of gap junctions between astrocytes has been connected with epilepsy. The aim of this study is to computationally test the influence of astrocytes connectivity in regulating activity in the neuronal network. To conduct the study, it has been used an in silico neuron-astrocyte model developed in our group. The model simulates the processes governing the communications between an astrocyte and a pre- and a postsynaptic neuron in the tripartite synapse, as well as between astrocytes through gap junction coupling. The modeled network comprises 250 neurons and 107 astrocytes. Three different astrocytic connectivity levels have been studied – representing 0, 2 and 4 gap junctions on average per astrocyte. Additionally, three different noise levels have been applied to the presynaptic terminal to simulate low, high and hyperactivity. Since the activation of astrocytes is driven by the activity of the neuronal network, the results showed that in case of low activity astrocytes were not activated and did not regulate neuronal activity. In case of high neuronal activity and hyperactivity, astrocytes showed an increased capability of downregulating neuronal activity when increasing the astrocytic connectivity. These results are in accordance with several in vivo experiments from different laboratories.

Barbara Genocchi, Kerstin Lenk, Jari Hyttinen
Heart Closed-Loop Model for the Assessment of Cardiac Pacing

We developed a closed-loop model of cardiac stimulation using a finite element model of the whole-heart embedded in the torso that is proposed as an useful tool for pacemaker design and testing. The electrical activity of the cardiac tissue is reproduced with a bidomain model incorporated with the FitzHugh-Nagumo equations. The finite element model is developed in Comsol Multiphysics, both in two and in three dimensions and then exported in Simulink environment where the pacemaker algorithm is implemented. To validate the model, we chose a demand inhibited pacemaker, which stimulates the myocardium only if the intrinsic activity of the heart is not revealed, but every type of pacemaker can be simulated. The model generates a controlled spontaneous activation in the sinoatrial node and it is also able to reproduce realistic electrocardiographic signals and the effects that the stimulation has on them.

Niccoló Biasi, Alessandro Tognetti
Model-Based Assessment of Sex Differences in Glucose Effectiveness and Its Components

Sex differences may assume a key role in condition of impaired glucose metabolism and progression to type 2 diabetes, affecting insulin-dependent processes. However, the presence of sex differences in non-insulin-dependent processes (i.e. glucose effectiveness) has been scarcely investigated. The aim of this study was to detect the presence of sex differences in glucose effectiveness (SG), as assessed by minimal model analysis, in subjects with different degrees of glucose metabolism impairment. Two groups of subjects ranging from normal (NGR, n = 57, males/females: 31/26) to abnormal glucose regulation (AGR, n = 115, males/females 42/73) underwent a 3-h frequently sampled intravenous glucose tolerance test. Minimal model analysis provided SG and its components at zero (GEZI) and at basal (BIE) insulin. Values for SG were 2.52 ± 0.98 10−2 min−1 and 2.81 ± 1.07 10−2 min−1 for males and females in the NGR group, and 2.08 ± 1.21 10−2 min−1 and 2.09 ± 0.98 10−2 min−1 for males and females in the AGR group. No statistically significant difference was found between males and females in both NGR (p = 0.29) and AGR (p = 0.94) groups. Sex differences were not detected for GEZI, which provided the major contribution to SG, either in NGR or AGR group. In conclusion, glucose effectiveness and its components seem to be not affected by sex differences in all glucose tolerance conditions.

Micaela Morettini, Ludovica Ilari, Christian Göbl, Alexandra Kautzky-Willer, Andrea Tura, Giovanni Pacini, Laura Burattini
Insulin Clearance in Women with a History of Gestational Diabetes Assessed by Mathematical Model Analyses of Intravenous Glucose Tolerance Test

Circulating concentrations of insulin are determined by a balance between the secretion rate of insulin from pancreatic beta-cells and insulin degradation (“clearance”). However, limited attention has been devoted to the study of insulin clearance in women with former gestational diabetes mellitus (GDM), which are known to be at increased type 2 diabetes risk. The aim of this study was to provide a detailed analysis of insulin clearance in women with former GDM. A population of 156 white Caucasian women, was analyzed early postpartum (4–6 months after delivery) and classified in two groups: women with previous GDM (pGDM, n = 115) and women that remain healthy during pregnancy (CNT, n = 41). All women underwent a 3-hour Insulin-Modified Intravenous Glucose Tolerance Test (IM-IVGTT). Insulin clearance temporal patterns were derived by mathematical modelling of IM-IVGTT data; average insulin clearance values were also considered during the whole test, and in the first - (0–10 min) and second phase (10–180 min). Insulin clearance temporal patterns were found to be different between CNT and pGDM group (p < 0.0001). Average insulin clearance was found different over the second phase of the test (p = 0.04), being equal to 0.54 [0.41] and 0.59 [0.41] l·min−1 in CNT and pGDM group, respectively. In conclusion, some abnormalities in former GDM women, compared to a group of healthy women were detected. This may be of relevance for more accurate estimation of type 2 diabetes risk.

Micaela Morettini, Christian Göbl, Alexandra Kautzky-Willer, Giovanni Pacini, Andrea Tura, Laura Burattini
Computational Models for Predicting Resilience Levels of Women with Breast Cancer

In the current study, a model-based system for predicting resilience in silico, as part of personalizing precision medicine, to better understand the needs for improved therapeutic protocols of each patient is proposed. The computational environment, which is currently under implementation within the BOUNCE EU project (“Predicting Effective Adaptation to Breast Cancer to Help Women to BOUNCE Back”), will help clinicians and care-givers to predict the patient’s resilience trajectory throughout cancer continuum. The overall proposed system architecture contributes to clinical outcomes and patient well-being by taking into consideration biological, social, environmental, occupational and lifestyle factors for resilience prediction in women with breast cancer. Supervised, semisupervised and unsupervised learning algorithms are adopted with the inherent ability to represent the time-varying behaviour of the underlying system which allows for a better representation of spatiotemporal input-output dependencies. The in silico resilience prediction approach accommodates numerous diverse factors contributing to multi-scale models of cancer, in order to better specify clinically useful aspects of recovery, treatment and intervention.

Konstantina Kourou, Haridimos Kondylakis, Lefteris Koumakis, Georgios C. Manikis, Kostas Marias, Manolis Tsiknakis, Panagiotis G. Simos, Evangelos Karademas, Dimitrios I. Fotiadis
A Systems Biology Approach to Decipher Genetic Variants in a Canine Model of Sudden Cardiac Death

Sudden cardiac death (SCD) represents a major public health challenge, accounting for approximately 25% of all cardiac deaths. It refers to an unexpected death from ventricular arrhythmia, occurring in individuals with preexisting cardiovascular disorders as well as in individuals not previously diagnosed with heart disease. The identification of genetic variants that increase susceptibility to SCD is fundamental to improve risk stratification and understanding of molecular pathophysiology. In this study, to investigate the molecular mechanisms underlying SCD, a canine model, recapitulating what may happen to patients with a prior myocardial infarction, was used to accomplish a genome-wide association study comparing dogs resulting susceptible or resistant to ventricular fibrillation during submaximal exercise. The identified variants were explored by means of a systems biology approach, which maps human orthologues of mutated genes into a network encompassing co-expression and physical interactions. The paths connecting mutated genes highlighted a subnetwork enriched for genes involved in regulation of cardiac function.

Martina Vescio, Lia Crotti, Peter Schwartz, Linda Pattini
Computational Fluid Dynamics Study of Inlet Velocity on Extrusion-Based Bioprinting

Extrusion bioprinting is one of the most studied additive manufacturing technologies thanks to the huge possibilities in the creation of engineered tissues that it provides. Despite all research performed about bioprinting there is not enough information on how inlet velocities affect the pressure distribution. In this way, we have carried out CFD simulations to check how different inlet velocities can affect the pressure distribution in the tip of a nozzle.

Juan Carlos Gómez-Blanco, Enrique Mancha-Sánchez, Juan Francisco Ortega-Morán, Antonio Díaz-Parralejo, Francisco Miguel Sánchez-Margallo, José Blas Pagador-Carrasco
A System to Assist in the Training of Medical Students in Respiratory Diseases

This work presents a model to train medical students in the analysis of different pathologies of the respiratory system. The model allows the simulation of dynamic behaviors of some variables associated with the pulmonary function under diverse physiopathological conditions. Variables such as age, gender, anthropometric factors and lung volumes obtained by plethysmography, can be used to personalize the model considering a particular patient. The model has been validated by comparing the results obtained by simulation with works by other authors. The two most common patterns of lung abnormality in clinical practice are presented as an example of applicability of the model, analyzing the relationships between inspiratory-expiratory flows and volumes, flow-volume loop and FEV1/FVC relationship.

Alejandro Talaminos-Barroso, Javier Reina-Tosina, Laura M. Roa, David Naranjo-Hernández, Gerardo Barbarov-Rostán, Pilar Cejudo-Ramos, Eduardo Márquez-Martín, Francisco Ortega-Ruiz
Effects of Arterial and Tracheal Pressures During a Respiratory Mechanics Protocol in Spontaneously Hypertensive Rats

This work aims to evaluate the response of tracheal and arterial pressures during a Methacholine (MCh) dose-response curve. This drug is a bronchoconstrictor and, additionally, causes vasodilatation in cardiovascular system leading to reduction of arterial pressure. Spontaneously hypertensive rats (SHR) were used and each animal was anesthetized, tracheostomized and connected to a small animal ventilator (flexiVent version 5.2, SCIREQ, Canada). Whereas the animal was ventilated, a device developed by us was connected to the right carotid in order to assess blood pressure. This device is composed of a pressure sensor, a low-pass filter, an amplifier and an analog-to-digital converter. The two main variables studied in this work were the arterial and tracheal pressure, the latter assessed by the ventilator. Firstly, we depicted the temporal dynamics of both pressures. The blood pressure decreased while the tracheal pressure increased with the increment of the MCh doses. In a second analysis, we compared the pressure during (intra) and after (inter) perturbations used to mathematical modeling. There were no differences between intra vs. inter, only among doses for the first three measures of each dose (p < 0.0001). Thus, despite the shape of the arterial pressure being different between intra vs inter perturbation, the mean value did not alter. Finally, we indicate that the arterial pressure returns to basal after 2 and 4 min after MCh injection, since that there was no statistical difference among doses or in 2 vs 4 min comparison (p > 0.05), which may guide the recovery time given between doses in a dose-response curve.

Amanda N. Barros, Vitor A. Takeuchi, Felipe Fava de Lima, Raissa R. S. Amorim, Otavio Henrique F. Ledesma, Maria Aparecida de Oliveira, Henrique T. Moriya, Renato Vitorasso
Modeling of Carbohydrates Oxidation Rate During Exercise in Type 1 Highly-Trained Diabetic Patients

Management of Type 1 Diabetes (T1D) in the context of exercise or sports competition still represents a great challenge for athletes living with this disease, due to the wide excursions in blood glucose level with increased risk of life threatening hypoglycemia. Recently, an algorithm called ECRES has been developed to estimate patient-exercise tailored carbohydrates (CHO) supplement required to maintain safe blood glucose levels during physical activity. This method estimates a CHO supplement based on the patient’s habitual therapy, the specific patient’s insulin sensitivity and the overall amount of CHO oxidized during the specific exercise. The last is based on the glucose pulse relation, i.e. the relation between heart rate (HR) and CHO oxidation rate, already studied in sedentary and moderately-trained subjects, but not in well-trained athletes. This study aimed to model the glucose pulse relation during exercise in type 1 highly trained diabetic patients and in healthy subjects. HR, oxygen consumption and carbon dioxide production were acquired breath-by-breath in seven T1D and seven well-matched healthy highly-trained subjects at four different exercise intensity levels, as well as at rest. Results showed a linear CHOox-HR relation (CHOox = 0.76 · %HRmax - 19.6; n = 70, R2 = 0.78) with no significant difference between the T1D and healthy athletes (p-value = 0.11). In conclusion, results of this study can be implemented in an updated version of the ECRES algorithm allowing an easy estimate of CHO supplement also in highly trained subjects. This useful support system can enhance the self-management of glycaemia during the training sessions of athletic patients throughout mHealth technologies.

Maria Pia Francescato, Miloš Ajčević, Alex Buoite Stella, Agostino Accardo

Regular Sessions: Biomechanics, Robotics and Rehabilitation

Pressurization of Axially Prestretched Tube: Consequences for Arterial Mechanics

It is well known that in their in situ position, arteries are axially prestretched. They retract upon excision as a consequence. Studies of the prestretch have shown that it depends on location in arterial tree. Axial prestretch increases with increasing distance from the heart. It has also been shown that the prestretch changes with age. Aging damages internal architecture of arteries and consequently longitudinal pretension is gradually loosen. Results obtained in laboratory experiments document that there is specific value of the prestretch under which axial deformation of the pressurized artery is negligible. Under such conditions, mechanical work performed by the pressure on axial displacements is negligible which is advantageous from biomechanical point of view. In the present study, by adopting the Gent model of the strain energy density function to characterize mechanical properties of the artery and assuming that this artery satisfies assumptions of the membrane theory, it will be shown that the value of prestretch, which minimizes axial displacement, also maximizes internal volume attained in the pressurization of the tube.

Zdeněk Petřivý, Lukáš Horný
A Closed-Loop Multiscale Model of the Cardiovascular System: Application to Heart Pacing and Open-Loop Response

A 1D description of the arterial tree is coupled to a lumped parameter model of the remaining circulatory system, resulting in a closed-loop multiscale model of the cardiovascular apparatus. The regulation of the arterial pressure is also implemented through a short-term baroreceptor model. The proposed framework reproduces well the physiological cardiovascular behaviour of an healthy young man and the modelled baroreflex mechanism is effective in adjusting the hemodynamic responses to both heart pacing and open-loop aortic-carotid sinus control.

Caterina Gallo, Luca Ridolfi, Stefania Scarsoglio
Experimental Study to Improve “Federica” Prosthetic Hand and Its Control System

Modern 3D printing technologies and wide availability of microcontroller boards allow to make active prosthetic devices in a simple way. This is the case of “Federica”, a very low-cost, under-actuated, active hand prosthesis. The five fingers of the prosthesis are moved by a single motor through inelastic tendons. The control system of the prosthesis is proportional to muscle contraction: firstly, EMG was used, then mechanical sensors that measure muscle volumetric variation were successfully utilized. This prosthesis proved to be particularly energy efficient and fast; it provided a general grasp function by adapting the exerted forces, thus allowing to easily catch even deformable objects. This study presents further analyses and design improvements of this prosthesis. In particular, a new, extremely simple but effective conditioning system of a force sensor resistor was presented and tested. In addition, the actual three-dimensional kinematics of a single finger was captured by means of high frame rate cameras and then analyzed. The new sensor conditioning system was characterized. It proved to be as effective as the EMG envelope to proportionally control the hand prosthesis motion, and it allowed an easier connection to common microcontroller boards. Kinematic analysis allowed to accurately reconstruct the actual phalanges motion over time.

Daniele Esposito, Chiara Cosenza, Gaetano Dario Gargiulo, Emilio Andreozzi, Vincenzo Niola, Antonio Fratini, Giovanni D’Addio, Paolo Bifulco
Study on the Activation Speed and the Energy Consumption of “Federica” Prosthetic Hand

Important features of a hand prosthesis are certainly the comfort in wearing it, the ease of use, the activation speed, the low energy consumption and no less important the anthropomorphic aspect. This study focused on the activation speed and the energy consumption of an under-actuated, low-cost, active hand prosthesis named “Federica”. The prosthesis is moved by a single servomotor able to rotate 180 degrees. Video acquisitions of complete rotations of the servomotor, when it works freely or fixed to the mechanical components of the prosthesis, were used to compare the different kinematic behaviors of the servomotor. A current sensor was used to measure the absorbed current, i.e. the energy absorption, by the servomotor under different uses of the prosthesis (at rest, grasping objects, raising water bottles, etc.). The comparison between the kinematic behaviors of the servomotor alone or connected to the prosthesis, showed the mechanical efficiency of the prosthesis with very low latencies and small variations in velocity and acceleration profiles. The prosthesis took about half a second from the muscle sensor trigger to the complete closure of the hand, showing a significant speed. Finally, tests on current absorption of the servomotor in various conditions resembling prosthesis daily usage, revealed the capacity to guarantee an autonomy of at least one day when powered by 7.4 V, 3000 mAh battery pack.

Daniele Esposito, Sergio Savino, Chiara Cosenza, Gaetano Dario Gargiulo, Antonio Fratini, Giuseppe Cesarelli, Paolo Bifulco
New Method to Analyze the Load Propagation on the Plantar Foot Surface During a Walk/Run Using the Smart Sock System

A new version of DAid® Pressure Sock System (DPSS) for gate monitoring and new method of data processing and analysis are presented. Proposed DPSS has 6 sensing elements, electronic device with sampling rate up to 200 Hz and is able to monitor walking/running gates in outdoor conditions. According to a new method the process of plantar loading during the gate is considered similar to seismic wave propagation and sensors assumed as the reference points. As a result, harvested data after being processed, are represented in the form of loading wave plots (WPs). Obtained WPs show load propagation across the sensors during the stance/gate. Primary tests of developed system and method were made for different types of walk/run and correspondence of conclusions from WPs analysis and diagnoses made by sport medicine doctor during walk/run tests confirms system efficacy and validity of results.

Alexander Okss, Alexei Katashev, Peteris Eizentals, Sandra Rozenstoka, Dace Suna
Intergame Analysis of Upper Limb Biomechanics of Stroke Patients in Real and Virtual Environment

The aim of this study was to realize an intergame analysis of upper limb biomechanics of stroke patients in real and virtual environment. Methods: The sample consisted of 11 hemiparetic patients, mean age of 51 ± 7 years. Participants made 15 attempts in two dart games (real and virtual). Elbow kinematics was video recorded during the dart throwing phase. Analysis was conducted using Kinovea software, paired Student’s t-test and Classification Regression Trees. Results: Patients exhibited a higher elbow extension angle (p = 0.008) and greater velocity in the real game (p = 0.005). In the virtual game patients had longer throwing time (p = 0.021) and better performance (fewer absolute errors) (p < 0.0001). The decision tree showed that there was a balance between the frequency of patients who played the virtual and real game and displayed elbow extension angles above 157°. Similar frequencies between velocity = 29 cm/s and >87 cm/s for the virtual and real games were found. In regard to dart throwing time, there was greater frequency of patients with time =1.37 s for the real game and >1.37 s for the virtual game. Conclusion: The patients can evolve satisfactorily in terms of angulation, velocity and time during virtual game training. Thus, we propose that the virtual dart game may be a useful tool in the neurorehabilitation of patients with chronic stroke, in line with therapeutic objectives and the patient’s clinical condition.

Herta Costa, Aline Fernandes, Débora Oliveira, Jamilson Brasileiro, Tatiana Ribeiro, Edgar Vieira, Tania Campos
The Effect of Perturbation Time on Selected Spatio-Temporal Parameters of Gait

The mechanisms that underlie suitable response in fall threatening situation are not well understood yet. Most of the knowledge has been obtained from slip and trip mimicking experiments and from subjecting individuals to external pushes. While most studies focused on balance responses where perturbations were delivered at foot strike in this study we examined how delaying perturbation onset to later stages of stance phase affect step length, step width and step time. The experiments were conducted on five healthy male subjects that were subjected to series of perturbation impulses of different amplitude, direction with respect to stance leg and onset delay. Results show that when perturbed in lateral direction with respect to stance leg healthy subjects increased step length and step time and decreased step width. However, the changes were less pronounced if the perturbation onset was delayed farther in stance phase. When perturbed in medial direction with respect to stance leg they decreased step length and step time and increased step width. However, the changes were less pronounced if the perturbation onset was delayed farther in stance phase. Results suggest that in healthy population perturbation onset plays important role in spatio-temporal responses.

Andrej Olenšek, Matjaž Zadravec, Zlatko Matjačić
Design of a Hybrid Portable System for Measuring the Position of the Spine, Pelvis and Center of Gravity of the Body

Posture and postural stability are important means of assessing the physical condition of a human. The system proposed in this paper allows to measure spine parameters (mutual position of individual vertebrae), mutual position of the pelvis and chest in the lateral plane, center of pressure (COP) under both feet and center of gravity (COG). It uses reflective markers as well as green markers to mark selected anatomical points that are captured by Microsoft Kinect 2 located behind the subject and Microsoft Kinect located to the subject’s side. COP and COG are determined using two Nintendo Wii Balance Board, stabilometric platforms which are placed next to each other. A MATLAB GUI application was created to retrieve data from the sensors, determine positions of the individual anatomical points and angles between them and the position of COP and COG. The experimental verification of the system has shown that the measured values, together with the determined statistical variables such as 95% confidence ellipse, can be used to assess the posture and postural stability.

Jan Hejda, Petr Volf, Monika Bačíková, Noa Bar, Cestmír Oberman, Kristýna Rusnáková, Marcela Braunová, Patrik Kutílek
The Evaluation of the Joint Quasi-Stiffness During the Robot-Assisted Gait Training: A Pilot Study

Walking may be defined as the forward displacement of the body requiring coordination between alternate successions of the swing phase and the stance phase. In literature it is known that, looking at the relationship between external moments and relative angles at lower limb joints during human walking is a new way of analyzing the biomechanical behavior of a joint (stiffness concept). Numerous studies show that the concept of quasi-stiffness applies particularly well to major loading phases of the lower extremity joints during stance phase. This study represents the first attempt to quantify the joint quasi-stiffness during the swing phase of gait over rehabilitation sessions performed by a child with cerebral palsy with the Lokomat system. The obtained results, albeit preliminarily, demonstrate the sensitivity of this new method in evaluating the degree of resistance offered by the patient over the rehabilitation sessions.

Luigi Iuppariello, Maurizio Nespoli, Fernanda Iammarone, Marianna Bertella, Ilaria Riccio, Marianna Cardillo, Angela Natalizio, Fabrizio Clemente, Mario Cesarelli
Design of Device for Measuring the Load of Cross-Country Ski Poles

Monitoring of physical activity allows the optimization of physical exercise and sports performance. Currently, there are a number of wearable devices designed to measure the acceleration of body segments. However, there are only a small number of experimental systems for evaluation of load and movement of cross-country skiing and Nordic Walking poles. The system proposed in the article allows to measure the force which user is applying on the pole, the pole linear and angular acceleration and the tilt of the pole. The system uses a pair of strain gauge bridges and inertial measurement unit mounted on each pole. The measured data are subsequently processed by the capturing and recording unit, transmitted via Wi-Fi interface and stored on an integrated SD card. Developed software allows analysis of the determined quantities including poling detection. The verification measurement has shown the usability of the system for assessing biomechanical parameters, particularly with regard to the evaluation of training methods and running styles for cross-country skiing and Nordic running. The limitation of the proposed system is the disruption of the structural integrity of the pole due to the bonding of the strain gauge bridges and the associated greater susceptibility to damage in the event of extreme load, e.g. in top sports.

Jan Hejda, Petr Volf, Jakub Mejstřík, Ján Hýbl, Aleš Tvrzník, David Gerych, Tomáš Michálek, Čestmír Oberman, Emil Bolek, Patrik Kutílek

Regular Sessions: Therapeutic and Diagnostic Systems, Devices and Technologies and Clinical Engineering

A Risk Stratification Model for Early Cognitive Impairment After Diagnosis of Parkinson’s Disease

Cognitive decline is very common in patients with Parkinson’s disease (PD) as longitudinal studies have shown that given enough time almost all patients will eventually end up with a diagnosis of Mild Cognitive Impairment (MCI) or dementia. Certain patients however are found to be more susceptible to early cognitive impairment soon after the diagnosis of PD. In this study, baseline evaluation outcomes from newly diagnosed patients from the Parkinson’s Progression Markers Initiative are evaluated to identify risk factors for early cognitive impairment using machine learning techniques. Applying an extensive search in the available feature space, consisting of more than 400 baseline features, to isolate the most informative predictors, the proposed methodology can discriminate patients having a diagnosis of MCI or dementia within the first 5 years of PD from those with normal cognition with an accuracy of 80.38%. Older age along with non-motor symptoms including cognitive and memory dysfunction, sleep problems, daytime sleepiness, smell dysfunction, mood impairment and anxiety at baseline are strong determinants of early MCI and dementia.

Kostas M. Tsiouris, Spiros Konitsiotis, Dimitrios D. Koutsouris, Dimitrios I. Fotiadis
Upper Limp Movement Analysis of Patients with Neuromuscular Disorders Using Data from a Novel Rehabilitation Gaming Platform

This paper describes the methodology for analyzing data from a novel platform that utilizes gamification techniques targeting patients with neuromuscular disorders that result in upper-limb movement limitations. The patient is asked to perform a number of sessions as prescribed by the physician for rehabilitation purposes, while the data are used to assess their progress. The hand movement is analyzed, and features are extracted regarding the movement patterns velocity, jitter, and trajectory. A set of sessions derived from healthy individuals have being recorded and analyzed so as to establish a baseline for the metrics. A statistical analysis of the differences between the healthy subjects and the patients is performed, helping us to focus on features of interest to the disease. The results will help determine how the patients’ motor skills are improving as the therapy progress, and accordingly adjust the number and type of sessions prescribed in a personalized manner.

Achilleas Chytas, Dimitris Fotopoulos, Vassilis Kilintzis, Theodoros Loizidis, Ioanna Chouvarda
3D Acquisition of the Ear Anatomy: A Low-Cost Set up Suitable for the Clinical Practice

The most common clinical treatment for ear deformities or non-congenital abnormalities is the reconstruction of the missing geometry using autologous costal cartilage. The surgical procedure consists in cutting, sculpting and suturing harvested costal cartilage from the patient to recreate an ear shape which is symmetric to the contralateral ear. During chirurgical operation, surgeons needs an accurate 3D template as reference to reproduce the ear. For this purpose, reverse engineering and additive manufacturing techniques can be employed. Specifically, this works aims to develop a reliable, low-cost and user-friendly system, to acquire the healthy ear geometry in clinical environment avoiding head patient’s exposition to radiation (MRI, CT scan). An ideal acquisition setup and device have been selected to achieve accurate results. To this end, a casted model of an ear was created as reference, and the best setup was evaluated by comparing the obtained 3D reconstructions with it. Once the setup has been determined, the anatomies of five volunteers were acquired, to test the methodology on human subjects.

Rocco Furferi, Elisa Mussi, Michaela Servi, Francesca Uccheddu, Yary Volpe, Flavio Facchini
Machine Learning Classification of Females Susceptibility to Visceral Fat Associated Diseases

The problem of classifying subjects into risk categories is a common challenge in medical research. Machine Learning (ML) methods are widely used in the areas of risk prediction and classification. The primary objective of these algorithms is to predict dichotomous responses (e.g. healthy/at risk) based on several features. Similarly to statistical inference models, also ML models are subject to the common problem of class imbalance. Therefore, they are affected by the majority class increasing the false negative rate.In this paper, we built and evaluated eighteen ML models classifying approximately 4300 female participants from the UK Biobank into three categorical risk statuses based on responses for the discretised visceral adipose tissue values from magnetic resonance imaging. We also examined the effect of sampling techniques on classification modelling when dealing with class imbalance.Results showed that the use of sampling techniques had a significant impact. They not only drove an improvement in predicting patients risk status, but also facilitated an increase in the information contained within each variable. Based on domain experts criteria, the three best models for classification were finally identified.These encouraging results will guide further developments of classification models for predicting visceral adipose tissue without the need for a costly scan.

Mahmoud Aldraimli, Daniele Soria, James Parkinson, Brandon Whitcher, E. Louise Thomas, Jimmy D. Bell, Thierry J. Chaussalet, Miriam V. Dwek
A Study on Relationship Between Walking Speed and Acceleration of Center of Mass Estimated with Inertial Sensors

Since the center of mass (CoM) is related to ground reaction force (GRF), the CoM position is expected to be used in evaluation of walking ability. The purpose of this study was to examine the possibilities of using the CoM acceleration calculated from the CoM trajectory estimated by inertial sensor signals to obtain similar information as the GRF. In this paper, 16 m straight walking measurements were performed under different walking speeds with healthy subjects. CoM acceleration, which was calculated from CoM positions estimated by signals measured with inertial sensors, were analyzed for the steady state gait. Waveforms of the CoM acceleration were similar among subjects. Four peak values (negative peak near the toe off (TO) and positive peak near the initial contact (IC) of the opposite lower limb in the anteroposterior component, and the positive peak before the TO of the opposite lower limb and the negative peak of the mid-stance phase) showed significant correlations with walking speed, which were similar to the relationships between GRF and walking speed in previous studies. These results suggest that anteroposterior and vertical components of CoM acceleration estimated from inertial sensor signals have a possibility of estimating similar information as the GRF during walking. The relationship between positive peak value of the vertical component near the IC of the opposite side and walking speed was seemed to be different from a previous study, although it showed significant correlation with walking speed. Simultaneous measurements with inertial sensors and GRF are expected in the next study.

Takashi Watanabe, Yuho Takeda
Comparative Assessment Between 3D and Conventional 2D Imaging Systems in Laparoscopic Practice

The objective of this study was to evaluate the effect of 3D imaging on laparoscopic suturing performance using a box trainer and during laparoscopic partial nephrectomy. Besides, the safety and usefulness of this technology in clinical practice were also subjectively evaluated. For the study in the training environment participated 30 surgeons and 13 surgeons in the study with the experimental animal model. Participants were organized in a random way into two groups, starting the trials with the 2D imaging system or with the 3D system. Three-dimensional imaging seemed to facilitate the performance of intracorporeal sutures in the box trainer setting. However, these improvements were not as clear during the partial nephrectomy. Surgeons stated that the 3D vision system helps improve depth perception during surgery and facilitates needle handling during intracorporeal suturing. However, they also indicated that its use can lead to the onset of visual discomfort or fatigue. Most of them would use again this vision system in surgical practice. Further studies are needed to more clearly determine the benefits of 3D over standard 2D imaging systems in laparoscopic practice.

Juan A. Sánchez-Margallo, Silvia Enciso Sanz, Francisco M. Sánchez-Margallo
Modeling of Transpalpebral Tonometry System for Parameters Optimization of the Measuring Sensor

This paper presents the method of intraocular pressure (IOP) determination by transpalpebral tonometry (TPT), especially for the diagnosis and the control of glaucoma. TPT can be preferred method for determining IOP in terms of minimizing the contact of the measuring device with the surface of the eye. The biomechanical basis of transpalpebral tonometry is briefly discussed. The two possible schemes of TPT technique which includes a stem with springs as measuring sensor are presented. Modeling of TPT system for parameters optimization of the measuring sensor in case the dynamic measurement method is described. The researches conducted with various combinations of stem weight and diameters are presented for IOP in the range from 10 to 30 mmHg.

P. V. Luzhnov, E. N. Iomdina, K. V. Ivanishchev, D. M. Shamaev, A. A. Kiseleva
A Feasibility Test of Evaluation of Gait Movement by Using Center of Mass Estimation with Inertial Sensors

It is necessary to control movement of center of mass (CoM) to realize stable walking. Since CoM has been shown to reflect asymmetry of movement, energy cost and ground reaction force (GRF), it is expected that CoM position makes possible to evaluate walking ability. This study aimed at examining the possibility of using CoM in gait evaluation, in which CoM position was estimated from signals measured only with inertial sensors. In this paper, 16 m straight walking of healthy subjects was measured with 7 wireless inertial sensors, and estimated CoM positions for the steady state gait were analyzed. Amplitude of change of vertical position of CoM during gait increased as walking speed increased in all subjects. Waveform of vertical position of CoM for each gait cycle was found to be divided into 2 patterns. The difference in the waveform of vertical position of CoM was considered to involve difference in gait movement to move CoM upward, in which knee joint angle or inclination angle of the shank segment was mainly changed for the term from the double stance phase to the single stance phase. This difference was considered to show the difference in gait strategy. The results of this study suggested that vertical component of CoM position became useful to evaluate gait movements.

Yuho Takeda, Takashi Watanabe
Controlled Thoracic Motions of an Anthropomorphic Phantom for Myocardial Perfusion Imaging

The aim of this study was to develop controlled thoracic motions on a cardiac phantom and a breathing phantom, which both are inserted within an anthropomorphic thorax phantom. The cardiac phantom consists of a left ventricle and the breathing phantom of inflatable lungs. The diastole and systole of the left ventricle follows the Wigger diagram while the inflatable lungs follow a sinusoidal exponential pattern. An ECG can gate both the beating left ventricle and the SPECT imaging modality. In addition, the ECG beating left ventricle follows the inferior portion of the inflatable lungs in the cranio-caudal direction at the level of the diaphragm. The lungs and the left ventricle are able to simulate normal and deep breathing conditions. A Programmable Logic Controller precisely controls these motions in time intervals of 10 ms. All motions can be operated simultaneously or independently to each other. A number of motion limitations of other physical phantoms have been overcome. This phantom assembly with variable cardiac defects can be utilized to optimize myocardial perfusion imaging.

Sotiris Panagi, Antonis Antoniou, Isabelle Chrysanthou-Baustert, Demetris Kaolis, Ourania Demetriadou, Costas Kyriacou, Yiannis Parpottas
3D Printing-Based Pediatric Trainer for Ultrasound-Guided Peripheral Venous Access

Peripheral venous access is an extremely common procedure, crucial in delivering drugs and collecting blood samples. It is associated to high failure rates, especially when pediatric subjects are involved, due to reduced limb size and low cooperation. Ultrasound can sensibly increase success rates and reduce the time required to perform the procedure, though a specific training is necessary to acquire adequate hand-eye coordination and simultaneously handle needle and probe. Commercially available simulators lack of realistic devices that reproduce anatomy and kinematics of pediatric patients. In this work, an echogenic simulator integrating direct 3D printing and silicone casting is proposed. More specifically, it replicates a five years old upper limb’s anatomy comprising an articulated skeleton, muscle tissues, skin and an integrated blood circuit. The devised simulator shows its effectiveness in terms of acoustic properties, articular kinematics reproduction and haptic feedback. Furthermore, the simulator can be easily customized according to specific training needs thanks to a highly flexible manufacturing process.

Rocco Furferi, Lorenzo Guariento, Kathleen S. McGreevy, Elisa Mussi, Niccolò Parri, Francesca Uccheddu, Yary Volpe
Pectus Excavatum: A New Approach for Monitoring Cup-Suction Treatment

The introduction of the vacuum bell (VB) for the conservative treatment of Pectus Excavatum (PE) has led to a new non-invasive alternative to thoracic surgery. The VB works by elevating the chest as long as a negative differential pressure is internally assured. In recent years studies have been conducted to validate this type of treatment and to outline its correct use; results show a short-term PE improvement when the device is worn for a minimum of 30 min (twice a day) up to a maximum of several hours a day for 12–15 months. Although the worldwide diffusion of VB devices increases year after year, its ability to lift the chest during treatment with respect to the applied pressure has begun to be evaluated only recently. In this paper, a new instrument for measuring chest elevation during treatment is presented and validated. The proposed system consists of two measurement devices: a commercial instrument for the detection of the negative pressure inside the VB, and a specifically developed optical system for the detection of chest movement. The effectiveness of the proposed system, tested on five paediatric patients, paves the way to the objective definition of an optimised patient specific VB scheme of use.

Francesco Buonamici, Antonio Marzola, Michaela Servi, Francesca Uccheddu, Yary Volpe, Marco Ghionzoli, Antonio Messineo
ARTE Project: EEG Analysis During Robotic Rehabilitation

The ARTE project has the objective of developing and testing an integrated rehabilitative system composed of a robotic end-effector, a wristband and a wireless EEG system. The main purpose of this integrated system is to monitor patients during the execution of a rehabilitation session to evaluate their performance and the effectiveness of their rehabilitation path. The system integrates the contribution of the Amadeo robotic end-effector (Tyromotion), the 8-channel EEG system Helmate (ab medica s.p.a) and the E4 wristband (Empatica). Starting from the EEG signal acquired during the rehabilitation session, indexes about activation of the motor cortex (ERD), level of engagement and brain asymmetry (BSI-Brain Symmetry Index) are computed. Moreover, the wristband extracts the patient’s heart rate (HR), inter-beat interval (IBI) and galvanic skin response (GSR) during the execution of the rehabilitation task: further information used to assess the subject’s condition and performance.The integrated device has been tested on a small group of healthy users. The obtained results were positive and demonstrated the feasibility of clinical trials.

Alessandra Calcagno, Stefania Coelli, Giulia Tacchino, Marta Baratto, Franco Molteni, Eleonora Guanziroli, Cosimo Puttilli, Anna Maria Bianchi
Bioimpedance, Total Body Water and Phase Angle of Preschool Czech Children: Preliminary Study

This study is a presentation of our ongoing process of obtaining body composition and electrical parameters from sufficient cohort of normal healthy children in Czech Republic that are currently not known. At the moment, 37 children have been fully measured and their parameters evaluated. The measured group consisted of 17 boys and 20 girls and their age was 5 to 6 years. Results are showing that hydration status and electrical parameters of children are approximately the same in the time of leave and in the time of arrival kindergarten. However, changes that are reported through all measurements indicate some specific behavior of bioelectrical values during children day cycle. Bioimpedance seems to act as a possible marker, that can describe in future body composition in another perspective than classical body composition parameters. Further work is needed to obtain sufficient cohort of normal healthy children and to be able to estimate new electrical parameters that may be used as reliable predictors of a health status of the child without using empirical equations or BMI.

Jan Hlubik, Lenka Vyslouzilová, Lenka Lhotská, Olga Stepankova, Jan Kriz
TOF-Watch NMB Monitoring Misleading Display Output During Moderate Neuromuscular Blockade

In general anesthesia, Neuromuscular Blockade (NMB) agents are administered to ensure intubation quality and complete immobilization. Monitoring NMB, based on the evoked response evaluation after a peripheral nerve stimulation, is essential to provide insight on medication dosing and suitable approaches for NMB reversal. Professionals often rely on the data displayed on the monitor screen after TOF stimuli, assuming measurements present the expected fade and a TOF-ratio > 90% is enough to rule out residual paralysis. This can be inaccurate and mislead the clinician to extubate before adequate NMB curarization. Data from 31 patients that underwent general anesthesia were retrospectively analyzed. All patients received a standard dose of rocuronium (0.6 mg/kg) for intubation, during maintenance additional rocuronium boluses were administered if solicited. NMB monitoring was done continuously applying TOF stimuli at the adductor pollicis with TOF-Watch SX®. Two types of monitoring display errors were studied: (1) valid TOF value without fade effect (invalid T1 > T2 > T3 > T4); (2) in the 30 min before extubation a valid TOF-ratio value > 90% (without error type 1) with T1 < 70%. Results show a mean (SD) of type 1 error on 42.0 (17.5)% of valid TOF measurements, type 2 error in 63 (45)%. Only 9 patients (29%) presented a TOF with no errors before extubation. These results may increase the distrust in the NMB monitoring devices, however its use should not be rejected when NMB agents are administered. Careful evaluation of the NMB is recommended and additional efforts should be placed in the accuracy of monitors data display to avoid errors.

Mafalda Couto, Catarina S. Nunes, Pedro Amorim, Joaquim Mendes
Device for Measuring Protection in Sunglasses Against Harmful Blue Light

There is a strong association between high energy blue light exposure and development of some ocular diseases such as age-related macular degenera-tion. In recent years there has been increased people’s concern about the damage caused by blue light and ways to prevent them. This study aims to develop and build a device for measuring protection against harmful blue light to human eyes in sunglasses and to determine whether this protection is adequate, adopting as criterion its blue light transmittance be less than 1.2 times its luminous transmittance. To obtain the proper spectral weighting functions, the measurement will consist of linear combinations of the responses of a TCS3472 photosensor illuminated by a high brightness white LED. The agreement between the measurements made with the proposed device and a gold standard will be analyzed by the Bland-Altman method. Sunglasses blue light protection values will be measured to analyze whether sunglasses in brazilian market are protected against harmful blue light. The developed device will be available to public to measure their own sunglasses.

Artur D. Loureiro, Liliane Ventura
Backscattered Ultrasound Periodicity Characterization on Trabecular Bone-Mimicking Phantoms: A Spectral and Wavelets Approach

In an attempt to provide better diagnostic techniques using quantitative ultrasound to assess bone’s microarchitecture (e.g. osteoporosis), this work aimed at characterizing the periodicity of trabecular bone-mimicking phantoms using backscattered ultrasound signals, by means of spectral methods and the continuous wavelet transform (CWT). Two transducers (3.5 and 7.5 MHz) were used on a pulse-echo configuration. Two bone-mimicking phantoms were studied, with different stiffness but similar mean scatterer spacing (MSS) distributions (varying from 1.5 to 2.5 mm). Transducers were translated parallel to the phantoms to collect signals for processing with the spectral autocorrelation (SAC), singular spectrum analysis (SSA), quadratic transformation (Simon’s method) and continuous wavelet transform for MSS estimation. Some methods were able to estimate MSS values inside the range informed by the factory. Differences in trabecular thickness led to differences in MSS estimates, mainly for 7.5 MHz. There is a possibility that spectral methods find more than one periodicity pattern.

Christiano Bittencourt Machado, Mahmoud Meziri, Wagner Coelho de Albuquerque Pereira, Guillermo Cortela
Short-Term Hemodynamic Variability in Supine and Tilted Position in Young Men

The aim of the study was to evaluate short-term changes in hemodynamic parameters observed in supine and tilted positions in six young men (age: 21–25). The cardiac inter-beat interval (RR), stroke volume (SV), ejection time (ET) and pre-ejection period (PEP) parameters were followed over two six-minute periods, in supine position and 10 min after a 60-degree head-up tilting maneuver, using continuously recorded impedance cardiography (ICG) and electrocardiography (ECG) signals. Hemodynamic variability was evaluated using coefficient of variation (CV), standard deviation (SD), and quartile deviation (QD). For the supine position, the mean (M), SD, CV and QD of the observed parameters were as follows. SV: 79 ml, 16 ml, 20%, 15 ml. RR: 905 ms, 106 ms, 12%, 163 ms. ET: 308 ms, 29 ms, 10%, 33 ms. PEP: 105 ms, 19 ms, 18%, 28 ms. In the tilted position, the following were observed. SV: 54 ml, 12 ml, 21%, 15 ml. RR: 706 ms, 68 ms, 10%, 68 ms. ET: 259 ms, 38 ms, 14%, 28 ms. PEP: 123 ms, 13 ms, 12%, 15 ms. The changes in hemodynamic variability caused by tilting were not unidirectional.

Gerard Cybulski, Edward Koźluk, Agnieszka Piątkowska, Ewa Michalak, Anna Stępniewska, Anna Gąsiorowska, Wiktor Niewiadomski
Efficacy of Machine Learning in Predicting the Kind of Delivery by Cardiotocography

It is well known that the interpretation of cardiotocographic (CTG) signals is still subjective and prone to misinterpretations; as a consequence, there has been an increase of cesarean sections, often not necessary, and initial expectations of significantly reducing perinatal morbidity and mortality have been unattended. Nevertheless, in developed countries, CTG is still the most widely employed prenatal technique for monitoring fetal health; and in many countries it represents a medical report with legal value. To overcome the drawbacks of the visual interpretation, many computerized systems for automatic or semi-automatic analysis of CTG have been developed in the last years. Recently, in order to support the diagnosis process, and to increase the predictive capability of these systems, also other techniques such as artificial neural network, deep learning and machine learning have been introduced. In previous works of the authors, software for automatic CTG analysis has been developed and described in detail. Now, by employing a dataset of features extracted from CTG signals with that software, to enhance its performances, different algorithms, such as J48, Adaboosting, Random Forests and Gradient Boosted Tree, have been tested to predict whether a birth would be a caesarean section or a vaginal delivery. The RF algorithm showed the best performance, since it reached the highest accuracy (87.6%), precision (87.9%) and AUCROC (93.0%). These preliminary results are very satisfying and encouraging; they confirm that to enrich the CTG analysis software with this methodology can help to significantly improve CTG classification.

Giovanni Improta, Carlo Ricciardi, Francesco Amato, Giovanni D’Addio, Mario Cesarelli, Maria Romano
Analysis of the Effect of Natural and Simulated Sun Exposure on Sunglasses Lenses: A Study on Materials Degradation

This work aimed to investigate the degradation of sunglasses lenses after a short period of exposure to solar radiation in two different ones. A set of 30 pairs of sunglasses lenses was tested, and the right lenses were exposed in an automatic sun exposure station, i.e. under natural conditions, and the left lenses were exposed in a solar simulator according to the parameters of the standard. In the solar station, the total exposure time was 115 h and in the solar simulator, 100 h. Lens spectroscopy measurements were taken before and after the exposure period for each experiment, and the UV, Visible and Infrared (IR) transmittances of the lenses were calculated by means of these measures, in addition to the accumulated erythemal UV radiation dose for both conditions. Exposure time was not sufficient to cause degradation of the lenses, but it was possible to verify some characteristics, such as differences between lenses of the same model, and differences between the conditions of the experiments that may influence the future results and the degradation of the lenses.

Leonardo Mariano Gomes, Mauro Masili, Liliane Ventura
Eye Scan Ultrasound System for Automatic Cataract Detection: From a Preclinical to a Clinical Prototype

Cataract is a pathology associated to the loss of the normal lens transparency, and its progression can result in a total loss of vision. The gold standard diagnostic method consists on qualitative observation through a slit lamp coupled to a microscope. This method has two limitations: incipient cataract may not be detected and the cataract hardness is subjectively evaluated. It can result on a late diagnosis and/or on surgical complications when phacoemulsification surgery (the most common therapeutic approach) is conducted. On this study we present a new technology for objective cataract characterization, based on ultrasounds. The Eye Scan Ultrasound System (ESUS) allows the characterization of cataract type, severity and hardness based on features extracted from A-scan signals, working at a nominal frequency of 20 MHz. A total of 27 features were extracted and analyzed through automatic classification algorithms. Preclinical and clinical prototypes of the ESUS have been implemented. From data collected on the preclinical phase, the system precision, sensitivity and specificity values were 99.7%, with a relative absolute error of 0.4%. Clinical phase is in progress.

Lorena Petrella, Marco Gomes, Fernando Perdigão, Mario Santos, Paulo Fernandes, Carlos Pinto, Sandrina Nunes, Miguel Morgado, Miguel Caixinha, Jaime Santos

Regular Sessions: Information Technology in Health Systems

Utilizing Incremental Learning for the Prediction of Disease Outcomes Across Distributed Clinical Data: A Framework and a Case Study

In this work, we highlight the need of a supervised learning framework for disease predictive modeling across distributed clinical data to overcome the privacy limitations that are introduced by centralized analysis. Towards this direction, a computational framework is proposed, consisting of six incremental learning algorithms that are based on Stochastic Gradient Descent, Naïve Bayes, and Gradient Boosting Trees, to provide new insight on the construction of supervised learning models across clinical data that are stored in multiple locations. The applicability of the proposed framework is demonstrated through a preliminary case study, where a distributed lymphoma prediction model is constructed across private cloud spaces that consist of clinical data from patients that have been diagnosed with primary Sjögren’s Syndrome (pSS). Our results reveal the dominance of the Gradient Boosting Trees, yielding an average accuracy 91.6% and sensitivity 87.5% towards the correct identification of lymphoma cases.

Vasileios C. Pezoulas, Themis P. Exarchos, Konstantina D. Kourou, Athanasios G. Tzioufas, Salvatore De Vita, Dimitrios I. Fotiadis
EmERGE Platform: A New mHealth Solution for People Living with HIV

The EmERGE platform is a novel mHealth solution that supports a new paradigm for HIV care. Following a rigorous co-design approach, it provides users with a system that can link securely into the health records of an individual, select the data required, and once reviewed by a clinician send the data and clinical opinion confidentially to the individual’s smart device. The proposed clinical pathway aims to facilitate patient empowerment and self-management of HIV while reducing face-to-face consultations for people living with stable HIV. The EmERGE platform has been deployed at 5 European clinical sites and it is being successfully used by more than 2000 individuals. A formal evaluation is currently underway using a tailored Health Technology Assessment process specifically developed for the assessment of mHealth solutions. The evaluation will assess the impact of the EmERGE pathway on patient empowerment, quality of life, patient related outcomes and experience measures, usability, quality of care and cost in diverse health systems.

Paloma Chausa, Francisco J. Gárate, Cesar Cáceres, Edward Wallitt, Jennifer Whetham, Enrique J. Gómez, the EmERGE Consortium
Machine Learning Algorithms Predict Body Mass Index Using Nonlinear Trimodal Regression Analysis from Computed Tomography Scans

In this study Machine Learning supervised regression and classification algorithms are used to predict Body Mass Index (BMI), starting from Computed Tomography scans (CT). From each patient CTs, 11 parameters describing muscle, connective tissue and fat, are extracted creating a patient specific soft tissue profile called Nonlinear Trimodal Regression Analysis (NTRA). Regression and classification are applied in order to predict and classify BMI using Tree-Based algorithms. A proper Train-Test division of the dataset is applied using k_fold Cross-Validation. Various combinations of features are employed with k_fold division in order to obtain the best coefficient of determination (R2) as evaluator of the quality of regression’s prediction. Afterward, BMI is divided into 3 and 5 classes and the same methodology is used to classify it. For this analysis, the accuracy parameter is calculated to evaluate the quality of the results. The max R2 is 0,83 and it is obtained using the 11 NTRA parameters as regressors, k_fold = 16, and the Gradient-Boosting Algorithm. The amplitude of the connective and fat tissue always covers more than 50% of all the feature importance. The best accuracy was 0,80 for 3 classes and 0,74 for 5 classes. The results prove that the 11 NTRA parameters can have a very significant predictive value and the same methodology can be applied in future works to predict other physiological parameters and comorbidities.

Marco Recenti, Carlo Ricciardi, Magnus Gìslason, Kyle Edmunds, Ugo Carraro, Paolo Gargiulo
Is It Possible to Predict Cardiac Death?

Cardiovascular diseases are the leading cause of death in all the world; despite having the knowledge of the main risk factors, they keep on being complicated pathologies to deal with. Cardiovascular management has introduced a lot of parameters as regards patients’ state of health; particularly, nuclear cardiology with Stress single-photon emission computed tomography myocardial perfusion imaging can carry out interesting parameters that have encouraged researchers to apply machine learning techniques to predict whether patients will die due to a cardiac event or not. The dataset consisted of 661 patients that were evaluated for suspected of known coronary artery disease at the Department of Advanced Biomedical Sciences of the University Hospital “Federico II” in Naples. Knime analytics platform was employed to implement a decision tree and Random forests. After a procedure of features reduction, 29 features were included, and the overall accuracy was 91.0%, while recall, precision, sensitivity and specificity overcame the value of 90.0%. This implementation shows the feasibility of machine learning combined with data coming from nuclear cardiology. Moreover, the possibility to predict cardiac death exploiting clinical data and parameters carried out from instrumental exams would help clinicians to provide patients with the best treatments and interventions.

Carlo Ricciardi, Valeria Cantoni, Roberta Green, Giovanni Improta, Mario Cesarelli
On the Privacy Enhancement of In-Transit Health Data Inspection: A Preliminary Study

A new healthcare paradigm has emerged lately placing patients at the center of the healthcare scenario, providing more control over their data. Allowing patients to access data from outside healthcare organizations may suppose a security risk and should be done cautiously. Perimeter security elements, such as firewalls, are placed at organizations boundaries to inspect the in-transit data, deciding whether it should be forwarded or not. However, a disjunctive appears, if health data is kept encrypted when traversing the firewall, it could not provide its functionality properly while revealing it could be a violation of the patient privacy. In this work we propose a new method based on multi party computation to perform a secure fine-grained deep packet inspection of health data and results are presented for different number of conditions and packets sizes when using the FHIR standard. Finally, we conclude that although the performance degradation introduced by the proposed method could be too high for a general use, it could be justified by the sensitivity of the involved data in healthcare scenarios.

Jorge Sancho, Gert Læssøe Mikkelsen, Jonas Lindstrøm, José García, Álvaro Alesanco
rOral: Use of a Teledentistry System for Remote Images Assessment in Oral Health Education Workflows

Medical training in oral health includes interpreting images of the oral cavity. The use of extended datasets, real cases, and the ability to monitor the progress of students is beneficial. In this work, we investigate the feasibility of using a teledentistry solution to support medical training. The rOral integrated teledentistry solution, developed by the research team, allows the decoupling of image acquisition and image review: images of the oral cavity can be recorded using the camera of regular mobile devices, near the subject of care; the images are then uploaded to a secure storage and reviewed by experts, using a web environment. This workflow, primarily designed for remote oral health diagnosis (e.g.: population screening), can be used for dentistry education, allowing students to play the role of experts. A population of dentistry students was involved in the present study and used rOral to access anonymized cases to form a diagnosis. The results show that the existing solution, planned for remote diagnosis, is suitable for education and that smartphone-acquired images can be used in diagnosis assessment activities.

Raquel Sebastião, Ilídio C. Oliveira, Ricardo Felgueiras, Nélio J. Veiga
Investigations on a Computer Application for Tracking the Mean Glandular Breast Dose Profile in Mammography

As quality assurance standards determine the use of single phantoms to be used in evaluation tests in mammography, checking the profile of the patients submitted to exams in each equipment is important to drive these tests. In addition, this profile is also relevant in investigations on the actual doses of radiation received by the patients during the exams, mainly in the current digital mammography systems. Therefore this work investigates these aspects regarding a single DR mammography unit, working in a public radiological service, by applying a novel software designed to extract data needed to the quality assurance evaluation from the DICOM files corresponding to the outcomes of the image exam. This software refers to a computational tracking scheme to investigate characteristics of the population submitted to the mammography exams, providing information on the dose profile yielded by the digital equipment. Data obtained by this application are compared to those provided by classical experimental tests and they were verified in agreement for such unit. They show that the most of patients relative to this study has a breast thickness profile above of that commonly used in test phantoms, being in the range of 52 to 70 mm. Furthermore the dose profile of this population was verified below the standard limits, which evidences the proposed software application as a useful tool to aid quality assurance programs in digital mammography.

Homero Schiabel, Bruno Barufaldi, Eny M. Ruberti Filha
Cuffless Blood Pressure Estimation Only an iPhone: Investigation on Cold Pressor Tests

Cuffless blood pressure (BP) measurement is an all-inclusive term for a method that aims to measure BP without using a cuff. In our previous article, we reported the development of a simple cuffless technique based on the haemodynamic expression, using only pulse rate (PR) and modified normalized pulse volume (mNPV) that can be measured using only a smartphone. The stressor used in our previous article was the cardiac-dominant pattern (mental arithmetic); however, here we have experimented with another stressor, namely, a vascular-dominant pattern in a cold pressor test, to verify this theory in 7 healthy young males and 6 healthy young females (N = 22). As a result, the mean arterial pressure accuracy was improved (r = 0.727) compared with that achieved in the previous study.

Ippei Harada, Noriyuki Mochizuki, Peter Rolfe, Masahiro Shibata, Takehiro Yamakoshi
The UBORA E-Infrastructure for Open Source Innovation in Medical Technology

The development of medical devices with open source and collaborative design methodologies has the potential to increase the access to medical technologies, thanks to a feasible reduction of design, management, maintenance, and repairing costs linked to the open access of device blueprints. UBORA is an e-infrastructure for the co-design of open source medical devices, which promotes the compliance with internationally recognized quality standards and regulations for safety and efficacy of devices, taking the EN ISO 13485:2016 and the EU MDR 2017/745 as inspiration. UBORA guides the user through a systematic design process, from the identification of clinical needs, of risks class and relevant standards for the device, and provides project management tools, including a repository, finalized to the preparation of the pre-production device dossier. The process is supervised by expert mentors, which ensure that safety and efficacy criteria are fulfilled. The UBORA e-infrastructure is in line with the 2030 Agenda for the Sustainable Development Goals, promoting and strengthening the initiatives of an international community of designers, healthcare providers and policy-makers, toward the reduction of inequalities in the access to medical devices.

Carmelo De Maria, Licia Di Pietro, Andres Diaz Lantada, Alice Ravizza, Mannan Mridha, Janno Torop, June Madete, Philippa Makobore, Arti Ahluwalia
Design and Implementation of a Web-Based Platform to Support Research in X-Ray Breast Imaging

This paper presents a web-based platform dedicated to evaluation studies in x-ray breast imaging. The backend development is based on ASP.NET and C#, while the frontend is implemented under HTML, CSS and JavaScript. The platform is designed to own simple and intuitive graphical user-friendly interface. Questions to be evaluated are entered by the users and stored in the application database. In the same database, two other features are kept as well: the results of the evaluation and the regions of interest taken from patient and simulated mammography images. The content of the database is subject to modification, as well as further addition of data. The designed web-based platform was tested in a particular evaluation study, which concerned the realism of the generated simulated mammograms from computational breast models. The new approach in evaluating images was estimated to be more convenient and faster compared to our previous experience. The system will be further updated and will be exploited in the routine research work of the laboratory as well as will be shared with other research groups.

Adelina Doycheva, Nikolay Dukov, Kristina Bliznakova
Empowering Diabetic Patients Using Gadgets and Mobile App

In this paper, we present our efforts towards more personalized diabetes treatment using the mobile application. The application helps to overcome the problem of missing and inaccurate data on blood glucose level, therapy, physical activity and other parameters, and allows development of the algorithms based on daily data of diabetic patients. For the purpose of our research, we developed two devices: myGluco - a small, portable device to be included into the glucose measurement kit patients carry with them on daily basis intended to transfer the saved blood glucose data from the glucose meter to the diabetic patients monitoring platform, and myWrist – the second small, portable device to track patients’ physical activity level and to serve in guiding the patients in performing strength exercises and to enable recording of their performance during exercising.

Sara Zulj, Goran Seketa, Dominik Dzaja, Luka Celic, Igor Lackovic, Ratko Magjarevic
A New Software Tool for Analyzing Mental Health Data in a Spanish Region

Mental Health disorders such as schizophrenia, depression, and dementia have a great impact on society worldwide. Recent developments have witnessed numerous advances in telemedicine that allow remote monitoring of elderly people with diseases of this pathology. One of key factors influencing recent advances in health information systems is the use of data extraction robust techniques used to extract knowledge from medical databases. The main objective of the paper is to develop a web application that allows extracting knowledge from a database of patients with psychiatric disorders and helps health personnel to know in depth the existing patient profiles and thus improve decision making. The scenario chosen for this study consists in a government of Castile and Leon database of 53641 income records of patients with Mental Health disorders between 2005 and 2015. Applying descriptive statistics for the analysis of data, the results show different parameters regarding Hospitalization behavior such as rate, average, hospitalization days of the patient’s number per hospital and outsourcing services.

Diego Calvo Barreno, Susel Góngora Alonso, Isabel de la Torre Díez, Miguel López Coronado, Manuel Franco
Feasibility of Machine Learning in Predicting Features Related to Congenital Nystagmus

Congenital nystagmus is an ocular-motor disease affecting people’s visual acuity since their first years of life. Electrooculography is used to perform eye tracking in these patients, giving the possibility to extract a wide variety of parameters. The relationships among all these variables were analysed in the past and the aim of this paper is to perform a new analysis employing more recent techniques, those of machine learning. The electrooculography of 20 patients was recorded, signals were pre-processed, and some parameters were extracted through a custom-made software. Knime analytics platform was chosen in order to build predictive models using Random Forests and Logistic Regression Tree algorithms and some evaluation metrics were computed. The visual acuity and the variability of eye positioning were predicted employing five and six variables, respectively. In terms of coefficient of determination, visual acuity had values over 0.72 and variability of eye positioning over 0.70. Compared to the results obtained without machine learning algorithms during the past years, these values become more valuable. In conclusion, this approach showed its feasibility in detecting relationships among variables related to congenital nystagmus; it could be tested in order to find new and stronger relationships among these variables and be of support for clinicians.

Giovanni D’Addio, Carlo Ricciardi, Giovanni Improta, Paolo Bifulco, Mario Cesarelli
A Smartphone Based Survey to Investigate the Cyber-Risk Perception on the Health-Care Professionals

A recent discussion has been opened on the cyber-risk in health care. Today the biomedical and the health informatics are under cyberattack. The study faces (a) from a general point of view the problem of the cybersecurity in the health care with particular attention to the medical devices of class 3 implantable active (which are vulnerable to the cyber-attacks) to the PACS, to the data and to the networks; (b) proposes and test a smartphone based survey to investigate the self-perception of the cybersecurity on the health-care actors.

Daniele Giansanti, Mauro Grigioni, Lisa Monoscalco, Rosario Alfio Gulino
ICT4MOMs: An ICT Integrated Approach to Monitor and Manage Pregnancy Development

ICT4MOMs project aims at implementing a novel remote ICT service towards the monitoring and prediction of maternal and fetal conditions throughout pregnancy. The envisioned application is based on the integration of wearable sensors and devices connected by an ad-hoc smartphone app in communication with an ob-gyn clinical center. Advanced signal and image processing software tools will be developed for extracting information from the recorded signals, namely: fetal heart rate, uterine contractions, continuous glucose sensors and portable US probe. Once validated by the clinical partners, the collected dataset will be used for a multivariate analysis based on soft computing classifiers and machine learning techniques. Based on the growing literature providing evidence on the fact that mother-fetus system should be considered as a whole, in this proposal pregnancy is conceptualized as a continuously evolving system which needs to be investigated by means of time-varying approaches. The crucial expected outcome is the integration of the established clinical knowledge with the results of computational analysis. Such multilevel integration is expected to provide reliable and translatable clinical guidelines towards a novel pregnancy management encompassing a more inclusive monitoring framework designed on a patient-specific level. The project was recently funded by the Italian Government—Progetti di Interesse Nazionale (PRIN) under the grant number 2017RR5EW3 for the duration of three years (2019–2021).

Maria G. Signorini, Nicolò Pini, Danilo Pani, Giovanni Magenes

Regular Sessions: Assistive Technologies

Smart Shirt for Uncontrolled Movement Retraining

Uncontrolled movement is a major cause of musculoskeletal pain and pathology. Retraining has proven to be an effective method for movement control improvement but it is a time-consuming process that requires much attention from a physiotherapist. The efficiency of movement retraining can be increased by independent work of a patient but it can prove difficult or potentially harmful if the patient lacks understanding of the correct movement or cannot verify the accuracy of the movement. This paper introduces a smart shirt system for movement control retraining assistance. The smart shirt system consists of 11 textile stretch sensors attached to a compression shirt for movement monitoring. To verify the ability of the shirt to assist uncontrolled movement retraining, a set of 12 tests were selected from the Kinetic Control method, and measurements were obtained both for correct and incorrect movements. The acquired measurements were analyzed to determine signals from sensors that would give the most significant information for each test depending on the involved anatomy. It was confirmed that the smart shirt system could be employed for evaluation of the test performance.

Peteris Eizentals, Alexei Katashev, Alexander Oks, Guna Semjonova
Computational Fluid Dynamics of Blood Flow at the Left Atrium and Left Atrium Appendage

In this study, we analyze local hemodynamics in the Left Atrium (LA) and Left Atrial Appendage (LAA) using Computational Fluid Dynamics (CFD). Our scope is to describe a novel application of CFD to analyze/describe the hemodynamic behavior at the LA and LAA according to different LAA anatomic morphologies and fluid velocities. 3D semi-automated reconstruction approach used to segment and reconstruct left atrium geometries and obtain the computational domain. Cardiac computed tomography scans (CTs) used from 3 different patients. Calculated velocities obtained from simulations varied significantly according to LAA morphology as follows: broccoli (0.1328 m/s), chicken wing (0.2468 m/s) and windsock (0.1268 m/s). We calculated as well the wall shear stress (WSS) at left atrium chamber as mean values 3 Pa, 5 Pa, 4 Pa for the patients, respectively.

Grigoris I. Grigoriadis, Antonis I. Sakellarios, Katerina Naka, Ioanna Kosmidou, Christopher Ellis, Lampros K. Michalis, Dimitrios I. Fotiadis
Powered Wheelchair Impact – User-Centered Observational Study

There is a growing prevalence of disability worldwide, which indicates an increasing number of persons who might benefit from assistive technologies. The purpose of this observational, descriptive, cross-sectional study is to assess the psychosocial and participation impact of powered wheelchairs (PW). From May to October 2017, 30 powered wheelchair users were interviewed using the Quebec User Evaluation of Satisfaction with Assistive Technology (QUEST) scale, the Psychosocial Impact of Assistive Devices Scale (PIADS) and the Activities and Participation Profile Related to Mobility (PAPM) scale, in addition to some demographic, clinical and powered wheelchair related questions. Descriptive and correlational statistics were performed to analyze the data. The participants were satisfied with both AT and related services, with the lowest QUEST scores belonging to those who had been using their wheelchairs for a longer period of time. PAPM scores revealed significant restrictions in participation mostly among users with longer wheelchair utilization. The most satisfied were the ones with better performance in terms of social participation. The psychosocial impact, in terms of adaptability, was higher among users who transitioned from a manual to a PW compared to those who already had a PW. There was an overall positive psychosocial impact of the powered wheelchairs, and, potentially, an increase in the quality of life of the users. There’s a lack of research in this area, specifically to evaluate the impact of the environmental barriers on the social participation and on the quality of life of powered wheelchair users.

Inês Domingues, João Pinheiro, João Silveira, Anabela Correia Martins
Virtual Assistant Prototype for Managing Medication Using Messaging Platforms

Healthcare chatbots (a.k.a. virtual assistants) offer services, tools, and interactions to the users, providing advice, help, support, and management to different diseases. The users of this type of chatbots can be, for example, patients, caregivers, and healthcare professionals, with different needs and requirements. Sometimes, patients with chronic diseases needing to take different medications several times per day, can have difficulties to remember their medication intakes. This situation can lead to negative effects on treatment effectiveness. In this paper, we present a virtual assistant prototype, aimed to help and manage patient’s medication through the use of Signal, a secure messaging platform that uses end-to-end encryption for better user privacy. The objective of this work is to create a handy tool to remind medication intakes through the creation of a supportive community where patients, caregivers, and healthcare professionals interact with helpful tools and services offered by the virtual assistant. Moreover, the developed prototype makes use of medical standards for data management and Artificial Intelligence Markup Language (AIML) for conversation support.

Surya Roca, Manuel Hernández, Jorge Sancho, José García, Álvaro Alesanco

Regular Sessions: Technologies for Active Ageing

“Patient Station” – Telerehabilitation System for People with Parkinson’s Disease

Parkinson’s disease is incurable and progressive impairment of movement. The physical activity combined with adequate exercises may improve motor capacity and delay the onset of activity limitation.European Physiotherapy Guideline for Parkinson’s Disease recommend strategies aimed to change behavior by facilitating intrinsic motivation. Tele-rehabilitation system which combines self-management of people with Parkinson’s disease (PWP) with ongoing supervision by a physiotherapist and/or neurologist can improve the quality of treatment of PWP and promote the individual approach of the therapist to the patient.In this paper, we describe the idea and realization of the core of such system, termed “Patient Station”, i.e. the platform to communicate between PWP and physiotherapist and/or neurologist. The system allows: (a) the assessment of the patient’s health state, (b) scheduling exercises, (c) assisting the patient when performing prescribed exercises, (d) controlling the execution of the schedule by PWP, (e) assessment the effectiveness of the rehabilitation, (f) supplying the documentation.The patient’s side of the system is based on smartphone application and inertial measurement units, communicating with a smartphone via Bluetooth connected with the therapist’s side by the internet-based database.

Marek Żyliński, Wiktor Niewiadomski, Aleksandra Wacławek, Aleksandra Budzyńska, Anna Gąsiorowska, Anna Stępniewska, Adam Becmer, Maciej Jagielski, Gerard Cybulski
An Overview of Assistive Robotics and Technologies for Elderly Care

Population ageing is already having an impact on societies. This study briefly reviews associated technologies employed within the framework of elderly care, namely: robotic nursing, ambient assisted living and assistive robotics. Their current status is considered together with their potential and the associated implementation challenges.

Eftychios G. Christoforou, Andreas S. Panayides, Sotiris Avgousti, Panicos Masouras, Constantinos S. Pattichis
Artificial Intelligence Gamified AAL Solution

GameAAL - Gamification Supporting Active and Assisted Living is an innovative system to monitor elderly daily life activities and provide neurocognitive stimulation games. The solution is focused at filling a gap on existing Ambient Assisted Living solutions, promoting social interaction and mobility of end-users using gamification and machine learning techniques. For each goal achieved, the system unlocks a reward providing the user with the chance of winning free admissions in cultural events supplied by community partners. A Proof of Concept of AAL ecosystem and TV interface is ongoing with 26 real end-users, reporting good acceptability and willingness to use the system.

Marta Pinto, Mário Pereira, Diana Raposo, Marco Simões, Miguel Castelo-Branco
Empowering Community Dwelling Older Citizens to Improve Their Balance with a Novel Technology Platform

The prevalence of balance deficits increases as the population is ageing. Such deficits are associated with the increased incidence of falls which in turn is linked with substantial limited functionality and morbidity. Vestibular rehabilitation therapy (VRT) as a component of the treatment has been shown to be effective in reducing symptoms and improving balance. HOLOBALANCE is an intervention based on a novel technology platform for providing VRT unsupervised, at home which means that motivating citizens to be compliant and promoting empowerment are the cornerstones for its wide adoption. Here we present how citizens empowerment is being addressed in HOLOBALANCE.

Dimitrios Gatsios, Doris Eva Bamiou, Sergi Costafreda, Eleni I. Georga, Konstantina K. Kourou, Themis Exarchos, Kostas M. Tsiouris, Dimitrios I. Fotiadis
Assessment of Tripping Hazards by a Single Step Evaluated by Principal Component Analysis of Pedestrian Feet Movements and Eye Behaviours

The aims in this study was to investigate feet movements and eye behaviours of young and older people when stepping over a small single step up to 25 mm height while walking. In addition, this study investigated what factors in feet movements and eye behaviours can be used to show effectiveness of the preventive measures of tripping. Healthy 12 young and 5 older participants joined this study. A walkway (9.6 m length × 1.2 m width) with a single step chosen from 0, 10, and 25 mm in one of three positions 4.8, 6, and 7.2 m from the start line, was used for measurement of feet movements and eye behaviours in stepping over. Principal component analysis demonstrated to evaluate pedestrian capability of stepping over the single step and showed differences between young and older groups. First principal component might show different walking style, second principal component might show foot strategy of stepping over the single step, and third principal component might show that awareness of the single step. The awareness level of 25 mm step by older group was similar level of the awareness of 10 mm by young group, and older group was struggle to aware 10 mm step, compared with 0 mm case. These findings might be useful to consider how to improve outdoor and indoor environments as well as to assess how physical training, preventive measures, and assistive technologies to reduce the risk of tripping in stepping over the single step.

Tatsuto Suzuki, I. Wa Liu, Nikolaos Papadosifos, Derrick Boampong, Pak Sum Fung, Nick Tyler

Regular Sessions: Biomedical Engineering Education and Society

Automatic Lung Reference Model

The lung cancer diagnosis is based on the search of lung nodules. Besides its characterization, it is also common to register the anatomical position of these findings. Even though computed-aided diagnosis systems tend to help in these tasks, there is still lacking a complete system that can qualitatively label the nodules in lung regions. In this way, this paper proposes an automatic lung reference model to facilitate the report of nodules between computed-aided diagnosis systems and the radiologist, and among radiologists. The model was applied to 115 computed tomography scans with manually and automatically segmented lobes, and the obtained sectors’ variability was evaluated. As the sectors average variability within lobes is less or equal to 0.14, the model can be a good way to promote the report of lung nodules.

Marlene Machado, Carlos A. Ferreira, João Pedrosa, Eduardo Negrão, João Rebelo, Patrícia Leitão, André S. Carvalho, Márcio C. Rodrigues, Isabel Ramos, António Cunha, Aurélio Campilho
Preliminary Validation of an Editable Virtual Reality Simulator for Minimally Invasive Surgical Training

MIS-SIM is a virtual reality (VR) environment designed and developed for the creation of virtual scenarios that can be used to train and acquire basic and advance laparoscopic skills. The environment is composed by a task editor where a content creator design and develop tasks for the simulator to play. Once they are completed, objective metrics are automatically stored and examined in MIS-SIM’s server so they can be displayed by an online platform. The project was validated in Semmelweis University in Budapest, Hungary where an experienced professor designed tasks for 16 young surgeons (PGY 3-4-5) from different surgical fields (gynaecology, general-, plastic-, vascular-, thoracic-, neurosurgery, etc.) with different experiences in laparoscopy. Each participant fulfilled each task as if they were completing them on physical simulator.

M. Rodríguez, D. Camba-Lamas, I. Oropesa, K. Juhos, L. Wauben, J. Dankelman, F. W. Jansen, G. Weber, E. J. Gómez, P. Sánchez-González

Regular Sessions: Clinical Engineering and Health Technology Assessment

Regulation and Approval of Continuous Non-invasive Blood-Pressure Monitoring Devices

New-generation devices that enable continuous non-invasive blood-pressure (cNIBP) monitoring as an alternative to invasive arterial blood-pressure monitoring are being developed. These include a volume-clamp method using transmural pressure control, a tonometer based on the principles of tonometry, and photoplethysmographic sensing calculated using either pulse transit time (PTT) or the second derivative signal. However, meta-analyses of cNIBPs indicate that they are not as accurate as intermittent cuff-based sphygmomanometer measurements. While most of the cNIBP devices have been approved by medical device authorities, regulatory bodies are only starting to discuss standards for them. This paper reviews the benefits and limitations of cNIBP devices and current trends in their regulation and approval.

Toshiyo Tamura
Evaluation of a New Endobronchial Double Lumen Tube with Integrated Camera: A Hospital Based HTA Experience

Over the last few decades, the developed countries have witnessed a strong increase in healthcare spending, essentially due to the emergence of new diseases, the aging of the population, and the development of new and costly technologies. The Health Technology Assessment (HTA) is a powerful tool for connecting the technical-scientific and decision-making world, which helps to prevent the introduction of inappropriate, ineffective, or superfluous technologies within the health system, thus limiting the expense, and improving the overall quality of medical care. This paper describes a hospital-based HTA procedure (costs-efficacy evaluation), which has been exploited to compare an innovative device, with integrated camera, for endobronchial intubation in thoracic surgery, with the instrumentation currently in use at the National Cancer Institute “G. Pascale” of Naples, which is the largest Institute of Hospitalization, Scientific Research and Care for the Oncology in Southern Italy. In the case under study, by introducing the new technology, direct variable costs could result in an actual significant spending increase for the “G. Pascale” Institute; however, despite that, the results of the HTA procedure and of the Analysis of Costs Minimization show that the introduction of the innovative device could allow to significantly decrease the intubation time, the overall surgery duration and the costs incurred by the Institute. These results confirm also the usefulness of the hospital based HTA procedures.

Michela D’Antò, Carlo Cosentino, Arturo Cuomo, Rossana Accardo, Paolo Bifulco, Leandro Donisi, Maria Romano
Design of an Evaluation Tool to Assess IoT Solutions for Active and Healthy Aging

Europe is currently the oldest continent having the “old-age dependency ratio” - persons over 65 compared to the active population. To cope with this situation, the deployment of Active and Healthy Aging (AHA) services through IoT is becoming increasingly relevant, to make this population as independent as possible within their environment. This study presents the whole process to design and implement the “Public Evaluation Website”, a tool that will present the results and achievements generated in ACTIVAGE, a Large-Scale Pilot that implements Internet of Things solutions for AHA in 9 European regions and cities.

Gloria Cea, Alba Gallego, Maria Teresa Arredondo, Giuseppe Fico
Practical Use of Early Stage Health Technology Assessment of Medical Devices: Systematic Literature Review

The early stage HTA (eHTA) is a commonly used tool for identifying economic indicators of health technology in early phases of research and development. It provides powerful influence instruments to manage the manufacturer’s investment risks and to improve the device design.There is no uniform generally accepted standard of eHTA implementation, but an analysis of several large-scale theoretical studies showed that, in general, the eHTA process for medical devices consists of three stages: conceptual modelling of the new medical technology, determination of new technology parameters by expert elicitation methods, and simulation allowing for an analysis of commercial options.The aim of this study is a systematic review of studies published from 2014 till 2019 and an analysis of selected articles in terms of the methods used in each step of the early stage health technology assessment.Most of the analyzed case studies deal only with a part of the early stage medical device assessment, probably because of its confidential character. But we have found the popularity of decision-making models, the applicability of AHP (the analytic hierarchy process) methods to the elicitation process, and information on the use of CEA (cost-effectiveness analysis) methods and headroom methods as economic analysis tools.

Mariia Simonova, Vladimír Rogalewicz, Gleb Donin, Peter Kneppo
Usefulness of the Blink Reflex to Assess the Effect of Propofol During Induction of Anesthesia in Surgical Patients

The aim of this study was to investigate the relation between the blink reflex evoked by an electrical stimulus and the depth of anesthesia induced with intravenous anesthetic drug propofol. The blink reflex was stimulated before the propofol infusion started (baseline) and after, every 6 s. The electromyographic responses and the level of sedation/anesthesia scores as well as the estimated effect-site concentration of propofol were recorded in 11 patients. The blink reflex responses were abolished when patients were still conscious. The clinical scale of anesthesia increased with increasing concentrations of propofol. To predict the level of sedation/anesthesia a multinomial logistic regression was performed using blink reflex extracted features at the frequency domain. Several features proved to be good predictor estimates and the model showed to be useful. This information could be helpful to assess the moment of loss of consciousness and thus personalize anesthesia.

Ana Leitão Ferreira, Catarina S. Nunes, Joaquim Gabriel Mendes, Pedro Amorim
A Novel Technique to Trigger High Beta and Low Gamma Activity in Patients with Schizophrenia

The main purpose of this study was to investigate the relationship between resting state and auditory-motor task electroencephalogram beta and gamma distribution in healthy subjects and subjects with schizophrenia. First, we looked to changes in the resting state EEG distribution in three healthy subjects and in three patients with schizophrenia. We also analyzed high-beta and gamma cortical activity from high-density EEG during a cognitively-driven auditory-motor task in two participants, one normal subject and one diagnosed with schizophrenia. For the auditory-motor task, we asked the participants to press a button using the thumb of both hands independently, during two three-minute sessions. Resting state EEG showed more fragmentation in schizophrenia patients when compared with the healthy participants. During the auditory-motor task, we observed increased cortical fragmentation and clustering during the hand response in the patient and healthy control compared with a resting state EEG. The fragmentation remains stable during both hand response time and reference period located between the command “press” and “do not press” suggesting that this task activates a network which remains similar during the entire task.

Eysteinn Ívarsson, Alec Shaw, Aníta Ósk Georgsdóttir, Brynja B. Magnúsdóttir, Aron D. Jónasson, Eric Wassermann, Paolo Gargiulo, Sigurjón B. Stefansson, Ovidiu C. Banea
P50 and P300 Event Related Potentials in Patients with Schizophrenia Recorded from High-Density EEG

The main objectives of this study were to describe P50 and P300 cortical topography in patients with schizophrenia and to define a robust methodology of signal quantification using high density EEG. Within a clinical trial in which patients with schizophrenia and auditory verbal hallucinations were submitted to 10 days rTMS treatment, P50 and P300 were investigated as possible neurophysiological measurements to assess treatment effectiveness together with psychometric scales like Psychotic symptom rating scale (PSYRATS), depression, anxiety and stress scale (DASS) and Quality of Life scale (QoL). Here we describe the technique of collecting P50 and P300 using high-density EEG and we reproduce the preliminary data of P300 in one healthy subject and P50 in two patients and two healthy subjects.

Ovidiu C. Banea, Elena Pegolo, Sara Marcu, Rún Friðriksdóttir, Eysteinn Ívarsson, Aron D. Jónasson, Viktor D. Jónasson, Brynja B. Magnusdóttir, Magnús Haraldsson, Eric Wassermann, Paolo Gargiulo
Total Cost of Ownership as a Management Tool for Medical Devices Planning: A Case Study of a ST-Analyzer in Perinatology

Total cost of ownership is studied as a part of Hospital-based HTA analyses for decision-making in the area of purchasing, implementation and/or disinvestment of technologies or interventions in hospitals. An estimation of Total cost of ownership for a D41 ST-analyzer for a neonatal department of a university hospital is presented as a case study. The choice of the time horizon and cost items to be involved is discussed. The necessity of discounting all cost items is accented. The method is fitting for estimation of life-cycle resource requirements in case of asset acquisition through donation or public subsidy.

Petra Hospodková, Petr Kudrna, Vladimír Rogalewicz
Cost-Effectiveness Analysis of Selected Methods of Haemostatis Evaluation

Viscoelastometric tests have a specific position in haemocoagulation techniques. Representatives of these methods are thromboelastography (TEG), and/or rotational thrombo-elastometry (ROTEM). These are fast bedside (POCT) methods based on a measurement of viscoelastic whole blood properties during the formation, stabilization and lysis of a blood clot. It was suggested that these methods could replace standard laboratory tests used up to now. The goal of this paper was to calculate and assess the cost-effectiveness in order to compare the viscoelastometric tests represented by the ROTEM delta device and standard laboratory tests (considered for a comparator) from the point of view of a healthcare provider. For this purpose, the purchase, maintenance and operation costs of both technologies were estimated, and the effects of both technologies were assessed using multiple-criteria decision-making methods. As a result, the use of viscoelastic tests proved to be the more cost-effective option.

Martin Zavadil, Michaela Blahýnková, Miroslav Selčan, Vladimír Rogalewicz

Regular Sessions: Neuro Engineering, Neuro Systems

Multimodal Approach for Epileptic Seizure Detection in Epilepsy Monitoring Units

Epilepsy is one of the most common neurological disorders, affecting up to 1% of the World population. Patients with epilepsy may suffer from severe consequences from seizures (e.g. injuries) when not monitored. Automatic seizure detection systems could mitigate this problem, improving seizure tracking and alerting a caregiver during a seizure. Existing unimodal solutions for seizure detection, based on electroencephalogram (EEG) and electrocardiogram (ECG) still have an unacceptable level of false positives, which can be reduced by combining these two biosignals. In this paper, EEG and ECG data from 7 epileptic patients with diverse recording length and seizure types were used for analyzing the importance of multimodal seizure detection, at a total of around 110 h 2 m. A leave one seizure out cross validation was selected, grouping data containing the period before a seizure and the seizure period. A proof of concept of multimodal seizure detection which uses a deep learning architecture directly on raw data is performed - a Fully Convolutional Neural Network and an architecture based on LSTM were tested. The network based on LSTM achieved better performance - using the best of one or a combination of both signals, all patients had above 91% detected seizures, a specificity per epoch above 0.96 ± 0.06 and a detection delay below 8.5 ± 12 s. These results show potential for developing a patient-specific approach for seizure detection that can be transferred to the ambulatory.

Paulo Maia, Elodie Lopes, Elisabeth Hartl, Christian Vollmar, Soheyl Noachtar, Joao Paulo Silva Cunha
Modulation of EEG Theta and Alpha Power by an Internal Attention Task with and Without Visual Distractors

Attention to internal representations is crucial in tasks involving temporary storage and mental manipulation of information (working memory tasks), to suppress concurrent interfering processes, such as sensory-intake processes. Theta activity (4–8 Hz), especially frontal, has been extensively associated to working memory tasks, increasing with task complexity and error monitoring. Alpha rhythm (8–14 Hz) has been hypothesized to enhance processes within the attentional focus via inhibition of task-irrelevant regions (alpha increase) and engagement of task-relevant regions (alpha decrease). Here, we provide further contribution to the investigation of alpha and theta activity by computing electroencephalographic (EEG) alpha and theta power from 10 participants who performed a resting phase (no attention) and a mental arithmetic task (internal attention), under two different visual conditions i.e., pictures and no picture presentation. Pictures stimulated external attention and acted as distractors during the math task but not during rest. Results show that frontal theta power increased during the math task vs rest, more largely in presence of visual distractors. Posterior alpha power increased in the math task vs rest in picture condition, but not in the no picture condition; moreover, it decreased in the picture vs no picture condition during rest but not during the math task. These results support frontal theta as a mechanism implementing cortical control in complex tasks and indicate that EEG alpha power reflects a sophisticate balance between increase in some regions and decrease in other regions, depending on the current goal and external context.

Elisa Magosso, Giulia Ricci, Mauro Ursino
EEG Motor Execution Decoding via Interpretable Sinc-Convolutional Neural Networks

The decoding of brain signals is a fundamental component of a brain-computer interface. Despite the success of deep convolutional neural networks (CNNs) in other fields, only recently these techniques have been applied to electroencephalographic (EEG) signals. One drawback of CNNs is the lack of interpretation of the learned features. In this study we introduce for the first time a sinc-convolutional layer into a CNN for EEG motor execution decoding, allowing a straightforward interpretation of the learned kernels. Furthermore, we apply a gradient-based analysis to assess the most relevant EEG bands for each movement and the most relevant EEG electrodes exploited in these bands. In addition to a slight accuracy improvement from 91.9 to 92.4%, our results suggest that the $$high\gamma $$ band is the most relevant EEG band, with gradient-based scalp distributions well localized at specific subsets of electrodes.

Davide Borra, Silvia Fantozzi, Elisa Magosso
Central Alpha Bicoherence Is Reduced in Photosensitive Subjects

The photic driving response is known to induce epileptiform activity in photosensitive subjects and a recent hypothesis suggests that a reduced inhibitory effect of the alpha oscillations may be associated to this phenomenon. In this work, we studied the linear and non-linear spectral characteristic of the physiological alpha rhythms during a visual stimulation protocol in a group of twelve healthy and eight treated photosensitive subjects. Results confirmed the desynchronization of the individual alpha power during stimulation over the occipital region in both the groups and over the central for the PS subjects. A reduced bicoherence of the central alpha rhythm in PS subjects was found, suggesting the presence of a different behavior in the cortical structures involved in the generation of the alpha oscillations, both at rest and during visual stimulation.

Stefania Coelli, Elisa Visani, Giulia Tacchino, Ferruccio Panzica, Silvana Franceschetti, Anna Maria Bianchi
Combined and Singular Effects of Action Observation and Motor Imagery Paradigms on Resting-State Sensorimotor Rhythms

In the present study, 30 right-handed participants randomly performed one of three motor neurorehabilitation paradigms: action observation (AO), motor imagery (MI) and combined action observation and motor imagery (AO+MI) of the right arm and hand movement. Resting state electroencephalography (EEG) was acquired for 5 min before and immediately after the motor paradigms session. EEG was recorded from 10 sites over sensorimotor areas, and the average power was calculated for left (FC3, C3, C1, C5, CP3) and right (FC4, C4, C2, C6, CP4) regions in the spectral bands: delta, theta, alpha, mu, low and high beta. Our main finding demonstrates that delta, theta and mu activity decreased significantly on the contralateral regions during MI, while low beta increased significantly. Except for the mu band, the same changes were observed on the ipsilateral side, where delta and theta decreased significantly, while low beta became significantly higher. No relevant effects were observed for AO or combined AO and MI. These findings demonstrate a rapid effect of MI on cortical modulation in sensorimotor areas which is revealed by changes in resting state oscillatory activity and suggest an interesting interplay between MI and AO. The presented findings may be relevant for choosing a proper protocol for clinical motor neurorehabilitation approaches.

Aleksandar Miladinović, Antonella Barbaro, Eddi Valvason, Miloš Ajčević, Agostino Accardo, Piero Paolo Battaglini, Joanna Jarmolowska
Network Analysis on Overnight EEG Spectrum to Assess Relationships Between Paediatric Sleep Apnoea and Cognition

One major risk of paediatric sleep apnoea-hypopnoea syndrome (SAHS) is the development of cognitive impairments among affected children. Electroencephalography (EEG) is ordinarily used as part of polysomnography, the standard diagnostic test for SAHS. However, how SAHS changes overnight EEG, and its relationships with cognitive performance, remains unclear. In this study, we first analyzed the spectral content of EEG recordings from 294 children to explore possible differences caused by SAHS. Then, a correlation network analysis was conducted to evaluate relationships among different EEG spectral bands and the results from the Differential Ability Scales (DAS) battery of cognitive tests. Our analyses identified up to six new SAHS spectral bands of interest in the EEG. They also showed higher absolute correlations among the different spectral bands as SAHS severity increased. Higher SAHS severity degree also indicated higher absolute correlations with DAS tests. Our results suggest that the spectral content of the overnight EEG is useful to characterize both paediatric SAHS and the cognitive performance of the affected children.

Gonzalo César Gutiérrez-Tobal, Javier Gomez-Pilar, Leila Kheirandish-Gozal, Adrián Martín-Montero, Jesús Poza, Daniel Álvarez, Félix del Campo, David Gozal, Roberto Hornero
Brain Processing During Postural Control – A Study Case

In humans, postural control is naturally unstable and is considered to be a complex motor skill derived from the interaction of multiple sensorimotor processes to maintain postural orientation. This is a case study based results from a previous research performed on a large cohort [1]. From a technological point of view, the aim is to characterize the cortical responses associated with abrupt balance perturbations (adaptation) and changes in cortical activity during prolonged postural control (habituation).. Also, to compare the cortical activity with the mechanical body sway performance during a balance task.Proprioceptive vibratory stimulation on calf muscles at 85 Hz was performed to evoke postural perturbation in a closed-eye experimental trial. Thirty seconds of quiet stance (baseline) and 225 s of random vibratory stimulation phases were divided into three epochs of 75 s each, with the impulses synchronized to EEG recording, using a 256-channel EEG cap and sampling rate of 1024 Hz. Channels shown to have statistically significant change in absolute power spectra variation were distinguished over six frequency bands (Δ, θ,α, β, low γ, and high γ). Force and torque actuated by the feet were recorded by a force platform and sampled at 50 Hz. A fluctuation index was computed to determine postural performance during the epochs.Results show a significant correlation between progress in platform performance and changes in cortical activity. The fluctuation indices decrease (indicating improvement in postural performance), and significant increase was found in absolute power values during adaptation phase in the parietal regions particularly in low gamma and thetaband.

Run Friðriksdóttir, Gunnar H. Karlsson, Halldor Á. Svansson, Fabio Barollo, Kyle J. Edmunds, Hannes Petersen, Paolo Gargiulo
Classifying Different Stages of Parkinson’s Disease Through Random Forests

Parkinson’s disease (PD) is a progressive, neurodegenerative and age-related disease whose clinical characteristics include both motor and non-motor symptoms. Gait analysis, a three dimensional, non-invasive and computerized analysis of gait, can analyse walking features and carry out spatial and temporal parameters that can be included in machine learning algorithms. Knime analytics platform is employed to implement Random Forests. The aim of the present research is to distinguish De Novo PD patients (patients in early phase, without treatment) and Stable PD patients (patients in intermediate phase, in stable treatment) using spatial and temporal parameters of gait analysis. The dataset consists of 59 people, 32.2% De Novo and 67.8% Stable patients. Results show high accuracy (84.6%) and capacity to detect De Novo patients (94.9% of sensitivity). Recall and precision got high values, too. Despite needing further investigation, this pilot research should encourage health policy and facilities to introduce machine learning techniques and gait analysis in clinical practice. Moreover, results suggest the existence of gait patterns characterizing each phase of Parkinson’s disease.

Carlo Ricciardi, Marianna Amboni, Chiara De Santis, Gianluca Ricciardelli, Giovanni Improta, Luigi Iuppariello, Giovanni D’Addio, Paolo Barone, Mario Cesarelli

Regular Sessions: Technologies for Preventive Healthcare

Photoplethysmogram Modeling of Extreme Bradycardia and Ventricular Tachycardia

A model for simulating life-threatening arrhythmias, namely, extreme bradycardia and ventricular tachycardia, in photoplethysmogram (PPG) signals is proposed. The model uses RR interval series as an input, thus publicly available ECG databases can be employed to generate PPG signal database. The model has been originally developed for simulating atrial fibrillation, however, is adjusted to better represent the variation in PPG pulse amplitude, intrinsic to life-threatening arrhythmias. The original and adjusted models are evaluated on 40 recordings containing the episodes of extreme-bradycardia and ventricular tachycardia. By using the adjusted model, the root mean square error between real and modeled PPG signals is reduced, on average, by 20.6% for extreme bradycardia, whereas remains similar for ventricular tachycardia. The simulator is expected to be valuable when developing PPG-based wearable devices for long-term monitoring of patients at high risk of sudden cardiac death due to life-threatening arrhythmias.

Birutė Paliakaitė, Andrius Petrėnas, Andrius Sološenko, Vaidotas Marozas
Evaluation in a Real Environment of a Trainable Cough Monitoring App for Smartphones

This paper presents SmartCough, an M-health app for Android smartphones that monitors cough trends in patients with respiratory diseases. The app is designed to be battery-efficient, fast, and robust against noise. It relies on efficiently-implemented machine learning algorithms that have been validated in laboratory conditions. Since these conditions are rarely met in a real situation where the user carries the phone inside their pocket or bag, the app features a self-training module that allows easy adaptation to new environments. In this paper, we have evaluated the app with real patients in an outdoor setting to test the performance in real environments that are hostile to cough detection. Our results show that the average sensitivity obtained in laboratory conditions drops significantly (down to 60%) when the baseline configuration is employed. By activating the built-in self-training module, the median sensitivity raises to 85.87% after a small training step, with a bounded false positive rate. The achieved performance is analogous to the one obtained in laboratory conditions, making the app suitable for use in real life scenarios.

Carlos Hoyos-Barceló, José Ramón Garmendia-Leiza, María Dolores Aguilar-García, Jesús Monge-Álvarez, Diego Asay Pérez-Alonso, Carlos Alberola-López, Pablo Casaseca-de-la-Higuera
New Approaches for Personalizing Daily Activity Monitoring in mHealth Applications

Healthy daily activities have a positive influence on many aspects of our lives. Habits have a deep impact on our health, they help to prevent the appearance of chronic and neurodegenerative diseases and will provide a healthy and active aging. This research work is aiming to analyze the need of new approaches on monitoring daily life activities, investigating new technologies and user modelling methods for healthy habits monitoring. mHealth platforms allow to perform a multivariable monitoring for allowing effective and personalized interventions. Data analytics, data mining and gamification methodologies are being applied to investigate user experience models. This user adaptation is commonly focused on personality and mood monitoring. Furthermore, new user models are built based on monitoring data, personality of user and the daily activity patterns extracted from intelligent and personalized monitoring. The final goal is contributing to improve user´s adherence to interventions, quality of life and quality of care in mHealth applications.

Diego Moreno-Blanco, Patricia Sánchez-González, Francisco J. Gárate, Cesar Cáceres, Javier Solana-Sánchez, José M. Tormos-Muñoz, Enrique J. Gómez
Encouraging Adherence of Chronic Obstructive Pulmonary Disease Patients to Physical Rehabilitation Programs Through Technology

Encouraging behaviour changes (e.g., healthy habits) is a recurrent problem addressed nowadays by adopting general interventions in hospital or primary care settings. However, professionals count with no time or resources to develop a long-term follow-up for each patient. The application of information technology (IT) on behaviour change intervention may bring a set of benefits such as personalized intervention, remote supervision and continuous accompaniment, increased efficiency, and optimization of human and material resources. Chronic obstructive pulmonary disease (COPD) is a worldwide current leading cause of death. Physical activity practice is one of the most effective non-pharmacological interventions for the improvement of COPD patients’ health state. However, adherence rate is low because performing home programs requires engaged and proactive patients or a continuous supervision by the professional. Both conditions are difficult to be addressed with current interventions.This work proposes addressing the high rate of COPD patients that drop out programs implying personalized rehabilitation at home through technology. This solution is approached by means of a controller that supervises a set of variables (collected from smart devices) and compares them with desirable outcomes in order to make a decision about adjusting the treatment, encouraging the patient or making amendments to the intervention. After assessing a first prototype by professionals and patients regarding usability of devices and preliminary interfaces, the current prototype is being piloted in controlled environments to check if it influences COPD patient adherence.

Jorge Calvillo-Arbizu, Laura M. Roa-Romero, Javier Reina-Tosina
Evaluation of an Environmental Autism Spectrum Disorder Monitoring Device

Currently, there is a growth in the diagnostic of Autism Spectrum Disorder (ASD) in children. This type of neurocognitive disorder prevents normal development from an early age. In addition, most people suffering from ASD have a Sensorial Processing Disorder (SPD) associated. This makes them perceive objects and environment in a different way and sometimes causes changes in behaviour, causing a crisis at certain times. It is proposed to monitor different physical parameters of the environment and the user to detect the causes that yield these changes in behaviour. For this, the paper shows the development of a first prototype that integrates some of those environmental parameters selected to measure and it is accompanied by a software for the management of the device. Finally, the correct operation of the prototype is validated.

José María Vicente-Samper, Carolina Blanco-Angulo, Ernesto Ávila-Navarro, José María Sabater-Navarro
Evaluation and Comparison of Text Classifiers to Develop a Depression Detection Service

Depression is a mental disorder which can become a serious health problem. This research line is focused on creating a depression detection service from text analysis. Sentiment Analysis and Natural Language Processing methods will be used to develop this service. The service will classify text in positive or negative depending on the emotions inferred from user’s input texts. In this study, five classifiers have been evaluated to determine which one fits better for this purpose. Naïve Bayes, Decision Tree, Naïve Bayes based on Bernoulli model and Maximum Entropy are the classifiers analyzed. A specific corpus which includes sentences that fit research needs was designed for testing purposes, whereas Sentiment140’s corpus was used with training purposes. An evaluation methodology formed by three tests was designed. Results show a promising starting point, but further analysis will be needed. Future related works will focus on expanding classification to get user’s mood and not only a binary classification. It will allow to include this service as input for developing personalized intervention and education mHealth systems.

Diego Moreno-Blanco, Borja Ochoa-Ferreras, Francisco J. Gárate, Javier Solana-Sánchez, Patricia Sánchez-González, Enrique J. Gómez

Regular Sessions: Biomaterials and Tissue Engineering

Co-encapsulation of Beta Cells and Nanoparticles Containing GLP-1 Greatly Improves Insulin Secretion in Alginate-Based Bioartificial Pancreas

The bioartificial pancreas presents a promising methodology to treat diabetes. However, this approach is still associated with a high rate of graft failure caused by poorly functioning cells. The incretin effect of glucagon like peptide-1 (GLP-1) make this peptide an attractive therapeutic agent to improve the success of transplantation of encapsulated beta cells. In present work, we developed a novel model based on the co-encapsulation of beta cells with nanoparticles containing GLP-1 on alginate hydrogels, allowing its delivery and action in the specific target, the beta cells.INS-1E beta cells were co-encapsulated with poly(lactide-co-glycolide) (PLGA) nanoparticles containing GLP-1 in alginate and were maintained in culture aiming to evaluate the performance of our drug-delivery system and the influence of GLP-1 in cellular performance. PLGA nanoparticles showed an average size of 169.3 nm, a polydispersity index of 0.05 and an average zeta-potential of −24.3 mV. The average association efficiency of GLP-1 was 65.4% and the in vitro release test showed 71% of GLP-1 delivery after 7 days. The presence of nanoparticles led to increased cellular metabolic activity and 8-fold higher insulin levels. Our results reveal a promising approach to ensure the proper delivery and action of GLP-1 in encapsulated beta cells, ensuring its viability.

Joana Crisóstomo, Francisca Araújo, Pedro Granja, Cristina Barrias, Bruno Sarmento, Raquel Seiça
Potentialities of LL37 for Wound Healing Applications: Study of Its Activity in Synergy with Biodegradable Composites Made of PVA and CA

Wound healing is a dynamic and complex process that results from the interaction between cytokines, growth factors, blood components and the extracellular matrix. Conventional dressings made of natural or synthetic materials have only the ability to manage the wound and protect it from repeated trauma. With the advancement of technology, wound dressings have evolved and are now capable of intervening in the healing process by targeting specific features of the wound, aside from protecting the wounded site. In this work, bioactive dressings capable of promoting healing and fighting infection in chronic wounds were explored. Various antimicrobial biomolecules were examined in light of their pathogen fighting skills and immunoregulatory potentialities. Dressing production processes were also investigated. Biodegradable composite dressings made of poly(vinyl alcohol), polycaprolactone, chitosan and cellulose blends were our main focus. The goal was to evaluate the synergistic effect of biomolecules and biodegradable polymeric dressings, considering the local and systemic treatment demands of chronic wounds.

Helena P. Felgueiras, Marta A. Teixeira, M. Teresa P. Amorim
Cellulose Acetate in Wound Dressings Formulations: Potentialities and Electrospinning Capability

Any open wound is a potential site for microorganisms’ invasion since their presence around us is inevitable. Skin wound healing relies on a series of complex physiochemical processes that remain a big challenge for healthcare professionals, particularly when the wounds are colonized by bacteria. Wound dressings play a major role in wound healing as they manage the wounded site, controlling the moisture balance and protecting the wound from repeated trauma, and by preventing possible infections from developing into more serious complications. Recently, bioactive dressings loaded with drugs and/or antimicrobial agents, allowing for a continuous and sustainable release of these molecules at the wounded site, have appeared in the market. Antimicrobial resistance is a growing health care problem, requiring more effective solutions than antibiotics. As such, nano- and microfibrous mats produced via electrospinning technique and loaded with natural-origin antimicrobial agents have attracted a lot of attention. Various polymers have been applied to engineer nanofibrous electrospun dressings. However, the environment impact of the synthesis and processing methods of synthetic polymers is undesirable. Therefore, the application of cellulose-derived materials (highly abundant polymer of natural-origin) becomes crucial as a green alternative to produce electrospun wound dressings with superior wettability, breathability and high capacity to promote cell proliferation, at relatively low costs. In this paper, different biomolecules loaded onto cellulose acetate (CA)-based polymeric nanofibers were investigated, and their antimicrobial properties were highlighted as alternatives to conventional antibiotics.

Marta A. Teixeira, M. Teresa P. Amorim, Helena P. Felgueiras
Electrospun Collagen Variability Characterized by Tensile Testing

The objective of this study was to verify repeatability of the preparation process of the collagen nanofibrous layers. The layers were fabricated by means of electrospinning. Constant conditions were held within the entire production process. It means that all samples were fabricated from collagen solution with the same chemical composition. Electrospinning conditions were also adjusted to have the same parameters, i.e. temperature, relative humidity, voltage and distance between electrodes. Finally, cross-linking procedure was also the same for all samples involved in the study. Elasticity of the material was assessed by means of uniaxial tensile test in a fully hydrated state of collagen samples. Internal structure of collagenous layers was studied by means of scanning electron microscopy. Although all technological conditions were held constant, mathematical modelling of the elastic behaviour revealed differences between samples. The results suggest that there is a certain variability of mechanical properties of electrospun hydrated collagen that is difficult to eliminate. By this fact, however, collagen is, to some extent, similar to biological tissue.

Ján Kužma, Lukáš Horný, Tomáš Suchý, Monika Šupová, Zbyněk Sucharda
Thermal Effect by Applying Laser Heating in Iron Oxide Nanoparticles Dissolved in Distilled Water

Due to their physical properties and biocompatibility, iron oxide nanoparticles have received particular attention in recent years for the localized hyperthermia therapy, where they are targeted to an organ, tissue or tumor and an external source is used for heating. Several physical, chemical and biological methods have been used to synthetize nanoparticles. A mechanical alloying method was used in this work to manufacture Fe2O3 nanoparticles. The size and shape of the nanoparticles were measured by scanning electron microscopy and X-ray diffraction. In this work, experiments were conducted with the nanoparticles dissolved in distilled water and heated by a laser in the near infrared range, with temperature measurements taken by an infrared camera. Numerical simulations were performed with COMSOL Multiphysics and compared to the experimental results. The numerical results agree with the measurements within the experimental uncertainties. The experimental results revealed a larger temperature increase of the sample surface for a larger concentration of nanoparticles. Hence, the Fe2O3 nanoparticles manufactured in this work behaved as a thermal agent with potential use for the hyperthermia therapy, including the treatment of cancer.

Leonardo A. Bermeo Varon, Bruna R. Loiola, Luiz A. da Silva Abreu, Bernard Lamien, Nilton Pereira da Silva, Helcio R. B. Orlande, Dilson Silva dos Santos

Special Sessions: Optimization in Medicine and Biology

Exact Linearization Techniques to Analyze the Population Dynamics of the Dengue Fever Vector

This work presents an analysis of a mathematical model describing the population dynamics of Aedes aegypti and the design of a vector control via exact linearization techniques and optimal control. The control strategy optimizes the use of resources for vector control. Theoretical and computational results are presented.

Helenice de Oliveira Florentino, Daniela Renata Cantane, Célia Aparecida dos Reis, Diego Cólon, Suélia Rodrigues Fleury Rosa
Advantage of Beam Angle Optimization in Head-and-Neck IMRT: Patient Specific Analysis

Radiation therapy (RT) main purpose is to eliminate, in a controlled way, all tumor cells sparing as much as possible the normal tissues. Intensity-Modulated Radiation Therapy (IMRT) is becoming the standard treatment technique in RT. Beam angle optimization (BAO) has potential to confer more quality to IMRT inverse planning process compared to manual trial and error approaches. In this study, the BAO advantages in head-and-neck patients are highlighted, using a patient specific analysis. Fluence optimization was done with Erasmus-iCycle multicriterial engine and BAO optimization was performed using two different algorithms: a combinatorial iterative algorithm and an algorithm based on a pattern search method. Plan assessment and comparison was performed with the graphical tool SPIDERplan. Among a set of forty studied nasopharynx cancer cases, three patients have been select for the specific analysis presented in this work. BAO presented plan quality improvements when beam angular optimized plans were compared with the equidistant beam angle solution and when plans based on non-coplanar beams geometries were compared with coplanar arrangements. Improvement in plan quality with a reduced number of beams was also achieved, in one case. For all cases, BAO generated plans with higher target coverage and better sparing of the normal tissues.

Tiago Ventura, Maria do Carmo Lopes, Humberto Rocha, Brigida da Costa Ferreira, Joana Dias
Determining Patient-Specific Dosage Scheme Using Integer Programming

Finding the correct substitution dosage of levothyroxine after total thyroidectomy can be time consuming using current methods. We present a mixed integer quadratic programming model for determining correct dosage given characteristics of a patient. The model is flexible in terms of determining fixed daily dosage regime, repeating schedule (e.g. weekly), and may include a loading dosage phase to achieve full substitution quicker. A similar model is used in an ongoing research project to determine dosages for patients. One challenge of the most sophisticated dosage regime is that it increases the burden for patients in following a dosage regime with frequent changes and where pills need to be divided to achieve more precise dosage.

Lars Hellemo, Vegard Heimly Brun
Optimization of Highly Noncoplanar Arc Therapy Trajectories: A Dosimetric Approach

The latest generation of linear accelerators allows the use of noncoplanar trajectories in arc therapy which combine the benefits of noncoplanar intensity-modulated radiation therapy (IMRT) treatment plans, such as improved organ sparing, with the benefits of arc therapy treatment plans, such as short treatment times. In this paper, we propose a two-step approach based on dosimetric criteria for the optimization of noncoplanar arc trajectories. In the first step, an initial set of anchor points (noncoplanar beam directions) is computed using a beam angle optimization (BAO) algorithm. In the second step, anchored in the beam directions already calculated, the noncoplanar arc trajectory is defined by iteratively computing additional anchor points considering the same dosimetric criteria used for the noncoplanar BAO. A nasopharyngeal tumor case already treated at the Portuguese Institute of Oncology of Coimbra (IPOC), is used to illustrate the benefits of the proposed optimization approach.

Humberto Rocha, Joana Dias, Tiago Ventura, Brígida Ferreira, Maria do Carmo Lopes
Dose-Response to Different Radiochemotherapy Regimens in Locally Advanced Pancreatic Cancer

Conformal radiation therapy (RT) delivered concomitantly with chemotherapy including 5-fluorouracil (5-FU) or Gemcitabine (GEM) is a common treatment for patients with unresectable locally advanced pancreatic tumors. In this study, the Poisson model describing tumor response to these two treatment options was derived. Clinical data was retrieved from reports published from 1990 to 2015. Dosimetric and clinical data from 1196 patients treated with RT with concurrent 5-FU or GEM were gathered. RT doses ranging from 3.6–64.8 Gy, delivered in fractions of 1.2–8 Gy, were converted to a 2 Gy fractionation scheme using the Biological Effective Dose concept. The parameters of the Poisson-Linear-Quadratic-Time model were derived using genetic algorithm optimization to minimize the least-square fitting error and a local search was then made using the maximum likelihood method. The goodness of the fit was assessed using the Pearson χ2-test. For RT+5-FU, D50 was 59.8 Gy, γ was 1.3, α/β was 3.2, Tpot was 18.6 days and Tk was 25.0 days. For RT+GEM, D50 was 54.5 Gy, γ was 1.4, α/β was 4.6, Tpot was 34.2 days and Tk was 37.2 days. As expected, RT+GEM showed higher efficacy than RT+5-FU. A RT dose-response effect was obtained showing that treatment strategies allowing a dose-escalation in pancreas tumors should be investigated.

Brígida C. Ferreira, Joana Dias, Adriana Gomes, Panayiotis Mavroidis, Humberto Rocha
Comparison of Different Radiotherapy Techniques for Locally Advanced Pancreatic Tumors

Radiotherapy (RT) associated with systemic therapy is the standard treatment for Locally Advanced Pancreatic Cancer (LAPC). The aim of this study was to compare the efficacy of different RT techniques using the clinical data reported in the literature. Clinical data was collected from scientific papers searched in the databases PubMed and ScienceDirect. Thirty-four documents published between 1997 and 2015 were found and met the inclusion criteria: locally advanced adenocarcinoma, unresectable and no metastasis. Values of Complete Response (CR), Partial Response (PR), Stable Disease (SD), Progression Disease, Progression Free Survival (PFS), and Overall Survival (OS) for Three-Dimensional Conformal Radiation Therapy (3DCRT), Intensity Modulated Radiation Therapy (IMRT) and Stereotactic Body Radiotherapy (SBRT) in the treatment of LAPC were collected. For all RT techniques, Response Rate (RR), defined as the sum of CR and PR, was for 3DCRT 25.2% ± 9.5 [range: 5.0%–49.0%], for IMRT 33.5% ± 10.5 [range: 10.6%–55.6%] and for SBRT 52.2% ± 17.7 [range: 13.3%–69.5%]. For all studied techniques, Local Control (LC), defined as the sum of RR and SD, ranged from 47% to 100%; PFS ranged from 4 to 12 months and OS ranged from 6 to 20 months. A significant improvement in overall response rate was obtained with SBRT compared to 3DRCT and IMRT. However, LC, PFS and OS were similar among the three RT techniques.

Adriana Gomes, Darlene Rodrigues, Brigida C. Ferreira
Optimal Location of Novel Robotic Prostrate Cancer Biopsy and Brachytherapy Treatment Devices

This paper details research towards the optimal placement of a novel robotic device for the detection and treatment of prostate cancer, via biopsy and brachytherapy respectively. A methodology for analysis of available data in order to determine geographical areas with relatively high prevalence of prostate cancer and low access to treatment is proposed. Areas in the South of the UK with high values of these indices are highlighted. The development of single and multiple criteria optimization models based on the new metric in order to optimally locate a small number of future prototype treatment devices is discussed. Discussions, conclusions and avenues for future research are given.

Sina Firouzy, Dylan Jones, Ashraf Labib

Special Sessions: Electronics and Smart Algorithms for the Effective Lung Monitoring and COPD Management

A Low-Cost USB-Compatible Electronic Stethoscope Unit for Multi-channel Lung Sound Acquisition

The diagnosis of respiratory diseases relies largely on auscultation. To further improve the diagnosis quality, continuous acoustic multi-channel surveillance of the airways is highly desirable. The application of multiple sensors around the thorax implies size and cost constraints for the sensors. In this article we first report on the design and development of a wearable and low-cost electronic stethoscope which is easy to integrate in a garment. We then analyzed the noise and linearity of the recorded sound signal and benchmarked a prototype with a standard stethoscope which is widely used in clinical practice.

Gürkan Yilmaz, Pierre Starkov, Mathilde Crettaz, Josias Wacker, Olivier Chételat

Special Sessions: Non-invasive Temperature Assessment Using Ultrasound

Effect of Continuous Application of Heating-Cooling Cycles on Ultrasonic Attenuation of Muscle Tissue

In ultrasound therapies with HITU applications, the temperature reached in the tissue depends both on the properties of the tissue to be treated and on the ultrasonic parameters. Ultrasonic attenuation (α) is one such parameters since it plays a fundamental role in the thermal source term of the biothermic equation. The study addresses the behavior of α by subjecting an ex vivo muscle to continuous heating-cooling cycles (15–50 °C). Using the transmission technique of ultrasonic pulse (PU) and a controlled thermal bath, the evolution of the pulses was obtained as a function of temperature. It was observed that, for the first three cycles, the evolution of the attenuation presents a non-reversible behavior. Then, the cycle becomes reversible, but at values very different from the initial value. Disregarding this behavior in the planning of a therapeutic session will result in an over or underestimation of the temperature, and, therefore, compromising the limits of safety and in the effectiveness of the application.

Guillermo Cortela, Carlos Negreira, Wagner C. A. Pereira
Metrological Approach for Characterizing Ultrasonic Properties of Soft Tissue-Mimicking Material

This paper reports the investigation of the metrological impact of temperature variation on ultrasonic velocity and attenuation coefficient of a standardized tissue-mimicking material (TMM), in temperatures ranging from 20 °C to 45 °C, in steps of 5.0 °C, using a new experimental method. One calibrated thermo-hygrometer was used to measure the temperatures of the water-bath in which the TMM was inserted. The pulse-echo technique was used to measure the TMM’s properties. The Guide to the Expression of Uncertainty in Measurement was used to estimate the measurements uncertainties. The experiments were performed under repeatability conditions (n = 4). Based on the results, it was observed that the values of the ultrasonic velocity tend to increase as the temperature increases, whilst the values of the attenuation coefficient tend to decrease. The group velocity varied from 1,533.7 m s−1 to 1,575.6 m s−1 with expanded uncertainties lower than 6.9 m s−1 (α = 0.05), for temperature varying from 19.9 °C to 45.8 °C. The attenuation coefficient varied from 2.55 dB cm−1 to 1.80 dB cm−1 with expanded uncertainties lower than 0.19 dB cm−1 (α = 0.05) for the same temperature range. The uncertainties found in this study are relevant to provide metrological reliability of the results since it takes into account the influence of the most relevant quantities involved in the measurement process for TMM characterization.

Raquel Monteiro Souza, Mylena K. Mosqueira de Assis, Rodrigo P. B. Costa-Félix, André Victor Alvarenga
Skin Contribution to Heating by Ultrasonic Field Irradiation: Simulation of a Multilayer Biological Application

Ultrasound diathermy is a very popular therapeutic method to treat musculoskeletal conditions. Nevertheless, there is no validated protocol and the physiotherapists empirically chose the dose to be applied. In this work, a computer simulation of therapeutic ultrasound propagated through a biological medium composed of four plane parallel layers representing skin-fat-muscle-bone was performed. The corresponding thermal field was obtained. Simulation was performed at 1 MHz, with 1.0 and 1.5 W . cm−2, with irradiation time of 120 s. with and without the skin layer (replacing skin by fat layer), to evaluate the role played by this layer. Results indicated that, when skin is present, it is the layer that reaches the highest temperatures, and the muscle layer did not reach temperature at therapeutic levels. When skin is excluded, then the fat/muscle and muscle/bone interfaces reach therapeutic temperatures in a very large region. These findings suggest that the skin cannot be neglected in dose planning. Additional experimental work has to be done to confirm this result.

Wagner Coelho de Albuquerque Pereira, Thaís Pionório Omena, Eduardo Moreno
Sensitivity Study in High Intensity Focused Ultrasound Therapy for Cancer

Hyperthermia using High Intensity Focused Ultrasound (HIFU) is an acoustic therapy used in clinical applications to destroy malignant tumors of bone, breast, brain, kidney, pancreas, prostate, rectum and testis. This technique consists in raising the temperature in the tumor or specific zone, to achieve coagulative necrosis and immediate cell death. For having a successful treatment, it is important to monitor and observe what is the tissue behavior, as well as its changes, before, during and after the procedure. In this paper, a sensitivity study is presented to determine the suitable parameters that may be estimated using mathematical modeling, thereby simulating an optimal treatment of cancer by heating induced by High Intensity Focused Ultrasound. The sensitivity analysis indicates the conditions under which temperature profiles are sensitive to changes in thermal conductivity and attenuation coefficient. This information provides the basis for estimation of the parameters in different tissues and for prediction of the thermal responses of these tissues.

Laura de Los Ríos Cárdenas, Leonardo A. Bermeo Varón, Wagner Coelho de Albuquerque Pereira
Improving Visual Contrast Between Fat and Muscle Tissues in B-Mode Images Using CBE: A Simulation Study

Ultrasonic radiation can be used as a non-invasive, ionizing radiation-free, portable and inexpensive imaging technique that allows to acquire images in real time. However, this technique presents some limitations, especially when referring to image quality, for example it is not easy to distinguish some soft tissues. Several methods for improving the visual quality have been developed, which are based, in most of cases, on disturbances of the medium with external stimuli. This work uses the changes in backscattered energy (CBE) to improve the visual difference between the muscle and fat tissue, because some studies show that CBEs calculated from RF signals present different behaviors when scattered by particles with different characteristics. The new image has more visual details when compare with the conventional ultrasound images. Two numerical phantoms were simulated with different kinds of scatterers. Conventional B-mode images were obtained for different temperatures (37 °–40 °C). A new parametric image was proposed by using the angular coefficient of the curve pixel intensity versus temperature, obtained for each pixel. The proposed parametric image was able to enhance the visual contrast between the simulated tissues.

Mario Pastrana-Chalco, Wagner C. A. Pereira, Cesar A. Teixeira

Special Sessions: Computational Biology and Medical Applications

CFD Analysis for the Evaluation of Patient-Specific Hemodynamic Parameters in Cerebral Aneurysms

Blood flow simulations are now considered a valuable tool for a deeper understanding of the physiopathology of intracranial aneurysms. Many authors built robust computational settings based on accurate computer-assisted registration, segmentation, and 3D geometry reconstruction from medical images of patient-specific cerebral aneurysms, and special techniques to derive appropriate boundary conditions. However, an accurate description of flow mechanics in the near-wall region and its connection with the evolution of the wall disease evolution remains linked to several questions not yet fully understood. Recently, several authors have suggested a lower order approximation of the Lagrangian dynamics in the near-wall region, which allows for a meaningful characterization of both normal and parallel flow direction to the wall. We verify this computational approach in a follow-up case study and try to provide a step further in the understanding of the hemodynamics environment and its possible connection with the aneurysm inception and growth.

Iolanda Velho, Jorge Tiago, Alberto Gambaruto, Adélia Sequeira, Ricardo Pereira
Wireless Capsule Endoscope Location and a Robotic Validation Experiment

We present results concerning the validation of a novel approach for wireless capsule endoscope localization, using as ground truth a simulated biological/mechanical environment experiment. The approach relies essentially on image-based methods. It involves a hybrid multi-scale affine and elastic image registration procedure which is afterwards appropriately complemented with calibration and visual odometry techniques. The capsule was fixed at the extremity of a robotic arm and moved along a part of an ex-vivo mammalian bowel. The first validation results indicate a good correlation between the ground truth velocity and distance traveled by the capsule and the velocity and distance given by the proposed approach.

Isabel N. Figueiredo, Luís Pinto, Luís Perdigoto, Marina Oliveira, Hélder Araújo, Pedro N. Figueiredo

Special Sessions: Smartphone Based, Patient-Centred Technologies

SmartBEAT: A Smartphone-Based Heart Failure Telemonitoring Solution

Heart failure (HF) is a chronic syndrome associated with poor prognosis and high healthcare costs, mainly due to frequent hospitalizations. These are preceded by changes in several physiological parameters, creating an opportunity for early detection and treatment of hemodynamic unbalance, and hospitalizations avoidance. Invasive telemonitoring systems have proven efficacy on this detection. Conversely, non-invasive systems need further evidence, mainly due to poor technological adherence.In this paper we describe SmartBEAT as smartphone-based HF non-invasive home telemonitoring system (NIHTS) and analyze preliminary results of SmartBEAT field test.SmartBEAT collects, analyses and manages HF patients’ data. Biological parameters are collected by external sensors and transmitted to a mobile phone and to a portal, where an algorithm identifies false from true alarms and displays a hierarchical interventional response.Patients were recruited in a Heart Failure Clinic at a University Hospital and used this system during 3 months, on a daily basis. Adherence to the system was analyzed and pre-/post-test data of usability assessments and perception of benefit were compared.A total of 10 patients aged ≥60 were recruited. Mean age 68 ± 4, 30% women, 8 ± 5 mean education years.Adherence to monitoring protocol was 97% at 3 months. Patients’ evaluation of this NIHTS was very positive in pre- and post-test assessments, reporting it was easy to use, conveyed a higher sensation of safety and was useful for healthcare management.

José Silva-Cardoso, Emília Moreira, Inês Lopes, Carla Sousa, Sérgio Leite, Manuel Campelo, José Maria Sousa, Manuela Fonseca, Linda Harnevo, Moshè Farin, Luís Filipe Azevedo, Filipe Sousa
How Secure Is Your Mobile Health?

Research suggests that the interactions between a patient and a health professional through an mHealth app (a mobile health application), can improve the efficiency and quality of healthcare. However, the risk of disclosure of personal or health related data can be higher when using such devices, as there is a lack of specific security standards or guidelines for their deployment. Also, there are commonly no controls to comprehensively verify and minimize security and privacy vulnerabilities that may exist in those apps before they are released in the “wild”. To make matters worse, the medical record has a significantly higher financial value (on the black market) compared with other personal records (e.g., credit card or bank account details), which obviously increases the motivation for unauthorised accesses and misuse. In order to mitigate these problems, socio-technical security as well as privacy and legal aspects need to be taken into account when developing mHealth apps. In this work, essential recommendations regarding the previously itemized are provided as a means to guide both developers and users alike, into more secure, private and usable mHealth apps.

Ana Ferreira, Rute Almeida, Joana Muchagata
Diabetes Management Guidance by a Logical Unit Supported by Data-Mining in a Mobile Application

Diabetes type I is a chronic disease that requires strict supervision. MyDiabetes is a utility application for diabetic users. This application served as basis to develop a logical unit, composed of logical rules, translated from medical protocols and guidelines, to advise the user. The data in the application is a source of knowledge about the user’s health state and diabetes intrinsic characteristics. An existing concern is the weak user adherence and consequential data absence. The implemented solutions were gamification and an interface rework. As later confirmed through a survey, users feel captivated by appealing interfaces, achievements and medals. In a near future, we will resume our work with the S. João’s hospital, with a new trial and volunteers. User testing will be used to validate the gamification techniques.

Diogo Machado, Vítor Santos Costa, Inês Dutra, Pedro Brandão
Automatic Quality Assessment of a Forced Expiratory Manoeuvre Acquired with the Tablet Microphone

Evaluation of lung function is central to the management of chronic obstructive respiratory diseases. It is typically evaluated with a spirometer by a specialized health professional, who ensures the correct execution of a forced expiratory manoeuvre (FEM). Audio recording of a FEM using a smart device embedded microphone can be used to self-monitor lung function between clinical visits. The challenge of microphone spirometry is to ensure the validity and reliability of the FEM, in the absence of a health professional. In particular, the absence of a mouthpiece may allow excessive mouth closure, leading to an incorrect manoeuvre. In this work, a strategy to automatically assess the correct execution of the FEM is proposed and validated. Using 498 FEM recordings, both specificity and sensitivity attained were above 90%. This method provides immediate feedback to the user, by grading the manoeuvre in a visual scale, promoting the repetition of the FEM when needed.

Rute Almeida, Bernardo Pinho, Cristina Jácome, João Fonseca Teixeira, Rita Amaral, Ivânia Gonçalves, Filipa Lopes, Ana Catarina Pinheiro, Tiago Jacinto, Cátia Paixão, Mariana Pereira, Alda Marques, João Almeida Fonseca
Combined Image-Based Approach for Monitoring the Adherence to Inhaled Medications

The adherence to inhaled controller medications is of critical importance to achieve good clinical results in patients with chronic respiratory diseases. To objectively verify the adherence, a detection tool was previously developed and integrated in the mobile application InspirerMundi, based on image processing methods. In this work, a new approach for enhanced adherence verification was developed. In a first phase template matching is employed to confirm the inhaler positioning and to locate the dose counter. In a second phase Google ML Kit framework is used for the detection of each numerical dose in the dose counter. The proposed approach was validated through a new detection tool pilot implementation, using a set of images collected by patients using the application in their daily life. Performance of each of the two phases was evaluated for a set of commonly used inhaler devices. Promising results were achieved showing the potential of mobile embedded sensors without the need for external devices.

Pedro Vieira-Marques, João Fonseca Teixeira, José Valente, Bernardo Pinho, Rui Guedes, Rute Almeida, Cristina Jácome, Ana Pereira, Tiago Jacinto, Rita Amaral, Ivânia Gonçalves, Ana Sá Sousa, Mariana Couto, Mariana Pereira, Manuel Magalhães, Diana Bordalo, Luís Nogueira Silva, J. Almeida Fonseca
Mobile Application to Support Children with Anxiety Disorders

This paper describes WEBECOOL, a hybrid web/mobile application with the purpose of improving the self-management of anxiety disorders. The development of WEBECOOL was driven by a multidisciplinary collaboration between computing engineers, psychologists and psychiatrics. The first stage of the web/mobile application test was performed to assess the satisfaction with the technological proposal. Also, several components were identified as crucial and were employed in WEBECOOL: stress management, problem solving, medication adherence, symptoms monitoring and social interaction. WEBECOOL also pretends to include caregivers in the process, with the purpose of implementing better help therapy strategies.

Nuno Fonseca, Ana Almeida, Maria Moreno, Raquel Simões de Almeida, Luiz Faria, António Marques, Paulo Matos, Pedro Rocha, Constantino Martins
Smartphone Recommendation System to Prevent Potential Injuries in Young Athletes

Over the last decades web and mobile technologies are increasingly being used in sports, especially in soccer, but none of them seems to allow to prevent injuries. However, training systems for young athletes do not have, for the most part, learning abilities in order to adapt, evolve and find new training recommendations. It is in this context which the Smart Coach project is presented in this work, and whose main goal is to introduce our mobile training recommendation system allow to young athletes evolve. The training mobile recommendation system is also designed to identify potential injuries risk for each young athlete.

Paulo Matos, João Rocha, Ramiro Gonçalves, Filipe Santos, Goreti Marreiros, Daniel Mota, Nuno Fonseca, Constantino Martins

Special Sessions: Computational and Experimental Modelling for Designing Bone Implant Systems

Experimental Assessment of Knee Arthrodesis

This study aims to evaluate experimentally the performance of the eternal fixator AO to promote an efficient knee arthrodesis, in terms of bone alignment. The knee arthrodesis serves as an option of salvage treatment for failed total knee arthroplasty procedures. Knee arthrodesis consists of the artificial formation of a joint between two bones (femur and tibia) through surgery, reducing pain, providing stability of the limb. It is also used to correct malformations or congenital deformations. In this work, the surgical procedure of knee arthrodesis was replicated by removing the knee joint, and having been applied compression loads in the femur and tibia, in order to promote bone fusion. The pressure distribution was evaluated in order to identify the occurrence of the bone fusion. The knee arthrodesis technique used was promoted with external fixation consisting on the union of two or more femoral pins with two or more tibial pins through sidebars. The objective is to verify the best bone alignment, in order to guarantee a good rate of bone fusion. It was possible to observe that the bone alignment presents great influence on the success of the knee arthrodesis, with an increase around 81% in the area of pressure distribution, when compared with the misaligned bones.

Ana M. Amaro, Luis Roseiro, Maria F. Paulino, Maria A. Neto
Implant-Assisted Removable Partial Dentures in Mandibular Kennedy Class I Patients: The Impact of Implant Positioning

The installation of two implants to support and retain the prosthesis has been suggested as an option to improve the prognosis of removable rehabilitations in Kennedy class I patients. Using the finite element method, the present study performed a biomechanical analysis of two implant-assisted options (IARPD) with the implants in distinct positions – mesially in the edentulous space or in a distal position. The installation of a dental implant in each of the bilateral edentulous regions promotes the vertical and anterior-posterior stabilization of the prosthesis, leading to low displacement. IARPDs with the implant placed adjacent to the abutment teeth retain a pattern of displacements comparable to conventional RPDs, with sinking of the prosthesis. Addition of implants to the denture bearing areas, regardless of the position, results in low strain in the edentulous bone and high strain in the peri-implant bone, indicating that load is distributed to the bone through the implants.

Ana Messias, Pedro Nicolau, Fernando Guerra, Ana Amaro, Luís Roseiro, Maria Augusta Neto
Residual Ridge Resorption in Mandibular Kennedy Class I Denture Wearers: Proposal of a Pressure-Induced Mechanism Based on a Finite Element Analysis

Removable partial dentures (RPDs) have been considered an acceptable cost-effective solution to restore oral function in partially edentulous patients. However, in mandibular Kennedy class I patients, the RPDs present unfavourable prognosis due to the combined tooth-mucous support with no posterior elements of retention and continuous resorption of the residual ridge. This process is inevitable but highly variable from one individual to the other and modified by both uncontrollable factors and functional or prosthetic factors that determine load distribution and might be subject of modification by the professional. Using the finite element method, the present study performed a biomechanical analysis of a mandibular distal extension removable partial denture regarding prosthesis displacements, strains in the supporting bone and pressure in the mucosa. The results suggest that bone resorption in the alveolar regions is a strain-mediated process whereas in the edentulous areas it is a process induced by pressure in the mucosa.

Ana Messias, Pedro Nicolau, Fernando Guerra, Ana Amaro, Luís Roseiro, Maria Augusta Neto
Numerical Evaluation of the Knee Arthrodesis Using a Modified External Fixator

This work presents the numerical results of arthrodesis using a modified bilateral-monoplanar external fixator for the treatment of septic sequelae of the knee joint. The knee arthrodesis serves as an option of salvage treatment for failed total knee arthroplasty procedures [1]. The arthrodesis technique used in this work was promoted with external fixation consisting on the union of two or more femoral pins with two or more tibial pins through bended lateral bars. Biplanar fixation has higher sagittal stability and higher fusion rates than the monoplanar fixation, but the level of compression on the fusion area as well as the extension of contact area can have a great impact on the bone union. Hence, the main purpose of this work is to give the surgeons additional information on how to influence positively the knee arthrodesis, namely relatively to the fixation level, the homogeneity of contact and to the compression level, but also about some possible complications. The arthrodesis technique was implemented using the CAD Solidworks® software and the numerical analysis was carried out on ADINA® software. With this methodology it was possible to recreate the mechanical procedure of the knee arthrodesis and the results indicate that the arthrodesis technique is not only influenced by the parameters associated with the kind of externatal fixator used, but also with the type of supported used to perform the external fixation assembly.

Maria A. Neto, Luis M. Roseiro, Maria F. Paulino, Ana M. Amaro
Finite Element Comparison of Two Implants for the Treatment of Unstable Trochanteric Femur Fractures

Fixation of unstable intertrochanteric fractures in osteoporotic bones often fails, due to lag screw cut-out. This study correlates the numerical results of a new Trochanteric Plate of Contention (TPC), which may improve the resistance to cut-out failure of internal fixation in osteoporotic bone, and the well-known Dynamic Hip Screw (DHS) system. Biomechanical properties of the new TPC met the ASTM F384-12 guideline requirements. Generally, it is well accepted that the DHS is the implant of choice in the treatment of stable intertrochanteric femur fractures [1], as well as is considered the implant that any new design should be compared with [2]. The histograms of the bone tissue level loading distribution, represented by the three principal strains in the femur head, were selected as comparative indices.

Maria A. Neto, Luis M. Roseiro, Maria F. Paulino, Ana M. Amaro

Special Sessions: Artificial Organs: Extracorporeal Blood Circulation Medical Devices

Computational Fluid Dynamics and Experimental Analysis of Blood Gas Transport in a Hollow Fiber Module

Membrane oxygenators or artificial lungs have become an important, reliable and lifesaving clinical technique. To limit side effects on blood platelet parameters and to reduce hemolysis the gas exchange in such devices has to be improved while the size of the membrane packing has to be reduced to meet geometric constraints. Computational fluid dynamics (CFD) provides a spatial and temporal resolution of the membrane oxygenation process and enables systematic optimization of artificial lungs. An innovative CFD approach was developed to examine the gas exchange performance of oxygenators. Blood and sweep gas flow in the fiber packing as well as blood gas exchange through the membrane between blood and sweep fluid were fully resolved and simulated. The results were compared to in vitro experiments comprising determination of blood side pressure loss and CO2 exchange performance of a prototype membrane module. This simulation approach provides a sound basis for the design of future artificial lungs.

Michael Harasek, Benjamin Lukitsch, Paul Ecker, Christoph Janeczek, Martin Elenkov, Margit Gföhler
Online Urea Concentration Estimation from Spent Dialysate Using Optical Sensor

The aim of this study was to estimate urea concentration on-line in the spent dialysate by a novel miniaturized optical sensor during hemodialysis and hemodiafiltration with different settings, and to compare the results with urea concentrations measured in laboratory. Ten end-stage kidney disease patients were enrolled for this study and the dialysate samples were collected, and signals were recorded with optical sensor during 5 midweek dialysis sessions with 5 different modalities for each patient. Using samples and optical signals from calibration set, the linear model was calculated to estimate urea concentration. For linear model the determination coefficient R2, the systematic error as BIAS and the standard error (SE) were 0.876, 0.00 ± 1.60 mM/L (calibration set) and 0.841, 0.01 ± 1.68 mM/L (validation set), respectively. In summary, the novel miniaturized optical sensor enables on-line estimation of urea concentrations during dialysis sessions and accuracy outperforms the measurement accuracy of UV–absorbance based urea concentration method from a previous study.

Kristjan Pilt, Jürgen Arund, Annika Adoberg, Liisi Leis, Merike Luman, Ivo Fridolin
Synthesis of Composites of Polyurethane Membranes/Polycaprolactone Fibers for Membrane Blood Oxygenators

Membrane blood oxygenators (MBOs) assure in extracorporeal blood circulation the metabolic functions of failing lungs. The technical and medical progress of the MBOs depends on two major factors: (1) hemocompatibility of the membrane/blood interfaces, and (2) enhancement of the flow management/mass transfer associated to the metabolic functions of the lung. Integral asymmetric bi-soft segment polyurethane (PEUU) membranes with polycaprolactone (PCL) as the second soft segment have been studied and showed enhanced hemocompatibility properties. In the present work, integral asymmetric PEUU membranes, containing 10%wt. of PCL (PEUU10), were synthetized by a modified phase inversion method. The synthesis of the composites, PEUU10/PCL, is carried out by the electrospinning of the PCL fibers over the top surface of the PEUU10 membranes dense layer. The electrospinning was performed in an in-house built set-up by applying an electric potential between the metallic syringe-tip with PCL solution and the target plate with a PEUU10 membrane. The solution was continuously fed to the syringe-tip at a constant flow rate and accelerated by the electric field to be deposited over the target. The PEUU10 membrane was attached to a cylindrical support target that allows to tune the fiber deposition conditions by varying the rotation speed. The PEUU10 membranes and the PEUU10/PCL composites were characterized by Scanning electron microscopy (SEM). The performance of the PEUU10/PCL composites in terms of boundary layer disruption, mixing promotion and subsequent oxygen mass transfer enhancement is evaluated in a surrogate system of a MBO.

Tiago Eusébio, Mónica Faria, Viriato Semião, Maria Norberta de Pinho
Hybrid Integral Asymmetric Cellulose Acetate/Silicon Dioxide Ultrafiltration Membranes for Uremic Blood Purification

Monophasic hybrid cellulose acetate/silica (CASiO2) integrally skinned membranes with silica contents between 5–18 wt.% were synthetized by an innovative method which combines the phase inversion and sol-gel techniques. The morphological and topographical characterization was performed by scanning electron microscopy (SEM) and atomic force microscopy (AFM). Permeation experiments were performed to determine the hydraulic permeability and rejection coefficients to reference solutes pertaining to the metabolic functions of the kidney. SEM confirmed asymmetric membrane cross-section structures and AFM showed that the introduction of silica reduced the submicron surface roughness at least 3 times compared to the pure CA membrane reaching a roughness mean value below 2.5 nm. Permeation studies show that the integration of silica into CA membranes increased hydraulic permeability of the hybrid CASiO2 membranes by a factor of ~2 and that all of the hybrid membranes fully permeate urea and totally reject albumin. In terms of hemocompatibility, all of the CASiO2 membranes are non-hemolytic, low thrombogenic and do note promote the highest stages of platelet activation.

Mónica Faria, Pedro Brogueira, Maria Norberta de Pinho

Special Sessions: Diabetes and Cardiovascular Diseases: Ibero-American Trends

Pulse Transition Time Method for Unobtrusive Blood Pressure Estimation

Pulse Transition Time (PTT) is being increasingly used for estimating blood pressure unobtrusively aiming at continuous assessment of cardiovascular diseases (CVD) and illness monitoring using portable and low-cost equipment. Several methods are available for PTT estimation, each one accepting its own constraints and physical parameters’ assumptions, besides the panoply of definitions attributed to PTT calculus. Defining PTT as the time difference between the time instant were peak R occurred in an electrocardiogram (ECG) cardiac cycle and the time instant were the photoplestimography (PPG) signal of correspondent cardiac cycle presents the inflexion point of its maximum slope, a method of PTT estimation based on measured ECG and PPG signals during several cardiac cycles is proposed. This method was tested on ECG and PPG data collected from 7 individuals for periods ranging from 30 min to 8 h. For the sake of PTT algorithm performance evaluation, ECG and PPG signals were collected with a commercial system, together with the PTT signals provided by the same equipment resultant from automatic computation with unknown algorithm. Preliminary results so far obtained encourage the use of the proposed PTT estimation algorithm on future internet of things (IoT) e-health system development aiming at elderly blood pressure estimation and CVD assessment. Accuracy of the proposed algorithm (0.9 correlation with reference signal) may be improved by further studies on raw data filtering parametrizations.

Maria G. Ruano, Amir Sadat Fazel, Ana Jiménez Martín, António Ruano, Juan Jesús García Domínguez
Improved Spectral Method to Obtain Strains of an Ex-Vivo Membrane Tissue and Its Performance Under Elevated SNRs

In the recent years, ultrasound has become an inspection tool under investigation to explore laminated biological structures to found mechanical or structural changes related to early stages of diseases. The resolution capability of ultrasound to explore these micro-changes not only depends on the frequency, but also in the processing methods applied to extract information from ultrasonic (US) signals. In this paper an analysis is presented to obtain strain values of the wall of a carotid artery segment (from an ex vivo porcine model) during a cyclic variation of the internal liquid pressure. Two methods were used to estimate the wall strain: the cross-correlation, which was considered as the gold-standard option up to 2018, and the SERHH spectral method which has previously demonstrated to obtain ultra-high resolution (0.1%) results on estimating temperature (2009) and wall thickness (2018) variations into biological tissues. Distinct results for the strain variations during the circulatory pressure cycle were detected with both methods (between 0 and 3.5 10−2 ± 0.6741 for the cross correlation and from 0 up to 3.0 10−2 ± 0.0488 for our SERHH method). This represents a considerable error of 17% if the cross-correlation options were used Moreover, a precise analysis for evaluating robustness of the SERHH method performance was performed under noised conditions. The original measured signals were added with different noise levels, and then new noise influenced strain values were predicted, for filtered and non-filtered noised signals. Results show that strains obtained for filtered signals almost maintain their original value, presenting small errors for the worst case.

Ivonne Bazán, Antonio Ramos, Carlos Negreira
Instrumental Proposal to Determine the State of Health of the Patients with Diabetic Foot

This work describes an instrument to make measurements of electrical impedance over the skin of the foot in Diabetes Mellitus patients, in order to provide an early detection of diabetic foot, a very severe complication of that disease. These measurements must be made in the area of the foot most prone to suffer ulceration in diabetic people. This electrical instrument is connected to a concentrator and distributor of multiple diagnostic information, whose purpose is to obtain and store the data from several measuring sub-systems related to a number of related physical parameters. This computerized global system permits the gradual expansion of the distinct foot parameters to be analyzed. Preliminary results indicate that there are significant differences in the measurements of electrical impedance in the very-low frequency frame, when diabetic vs. non-diabetic curves are compared.

Ilse Anahi Torres, Lorenzo Leija, Arturo Vera, Josefina Gutiérrez, Antonio Ramos
A CYTED Network: New Non-invasive Ways for an Early Diagnosis of Chronic and Degenerative Diseases: Diabetes and Cardiovascular

In this work, an analysis of the principal characteristics, scientific aims and some initial results, of the Iberoamerican R&D network “Ditecrod” (Project of CYTED) is made. This network is based on an international agreement to propitiate an efficient trans-disciplinary cooperation in a scientific area where currently significant research efforts are being made for their high potential impact on future health and quality of life: The non-invasive Early medical diagnosis of Cardiovascular and Diabetic Foot diseases. Both chronic & degenerative pathologies are endemic today in many American countries. The main project objectives are the research, application and diffusion of new non-invasive diagnostic methods, having a low technologic cost and an easy portability, in relation to nowadays disposable approaches. The main results (in the first year of operation) of this multinational network are described.This project also seeks to promote the cooperation working among 11 working teams integrated in six R&D groups with large experience and previous innovations on the subject, and 2 emerging university groups until now focused mainly on the university teaching; in total 13 teams of 8 countries: 5 of Bio-medical Engineering, 4 of Biophysics, 2 hospitals and 2 companies.

Antonio Ramos, Lorenzo Leija, Carlos Negreira, Eduardo Moreno, M. G. Ruano, Wagner Coelho, Ivonne Bazán, Fernando Merchan, César Yegros, Juan Prohias
Computational Strategy for the Generation of the Clinical Histories of Patients with Diabetic Foot

In patients, suffering diabetes mellitus by long time, the pathology of the diabetic foot is a consequence of that serious disease, appearing in many cases. This problem can be prevented through the evaluation and tracking of certain physical characteristics into tissues composing the patient feet, in such a way that, specific therapies could be applied. For instance, the blood flow in some foot zones is correlated with the tissues health status. Moreover, it would be possible to follow the effectiveness of some applicable preventive therapies, through a protocol based on regular clinical measurements. Here we explain the strategy used to generate a computer system, designed to dynamically create and maintain the medical records of patients with diabetic foot. The system can be used through a Web site, specialized in diabetic foot, including clinical and researching institutions having the possibility of sharing a complete information about many patients with diabetes: (i) data logging about patient medical histories, (ii) medical diagnostic observations, (iii) distinct therapeutic prescriptions, (iv) particular notes of the treating medical specialist, (v) and to follow-up the evolution of ulcers or predictive changes in the areas most prone to ulcerations. This powerful computational tool was designed for an easy handling, and an efficient inter-communication of a main server with a number of user computers connected through a medical network. Some results for “diabetic foot” tracking are shown. From its use by medical specialists and their successive observations, adjustments in the initial programming and possible extensions of this clinical tool are expected in the future .

Ilse Anahi Torres, Lorenzo Leija, Arturo Vera, Josefina Gutiérrez, Antonio Ramos

Special Sessions: Smart Robotic Assistant for Minimally Invasive Surgery: The SMARTsurg Project Experience

Towards Finger Motion Tracking and Analyses for Cardiac Surgery

Robot Assisted Surgery is attracting increasing amount of attention as it offers numerous benefits to patients as well as surgeons. Heart surgery requires a high level of precision and dexterity, in contrast to other surgical specialties. Robot assisted heart surgery is not as widely performed due to numerous reasons including a lack of appropriate and intuitive surgical interfaces to control minimally invasive surgical tools. In this paper, finger motion of the surgeon is analyzed during cardiac surgery tasks on an ex-vivo animal model with the purpose of designing a more intuitive master console. First, a custom finger tracking system is developed using IMU sensors, which is lightweight and comfortable enough to allow free movement of the surgeon’s fingers/hands while using instruments. The proposed system tracks finger joint angles and fingertip positions for three involved fingers (thumb, index, middle). Accuracy of the IMU sensors has been evaluated using an optical tracking system (Polaris, NDI). Finger motion of the cardiac surgeon while using a Castroviejo instrument is studied in suturing and knotting scenarios. The results show that PIP and MCP joints have larger Range Of Motion (ROM), and faster rate of change compared to other finger/thumb joints, while thumb has the largest Fingertip WorkSpace (FWS) of all three digits.

Mohammad Fattahi Sani, Sajeeva Abeywardena, Efi Psomopoulou, Raimondo Ascione, Sanja Dogramadzi
Surgeon Training with Haptic Devices for Computer and Robot Assisted Surgery: An Experimental Study

Development of robot-assisted minimally invasive surgery increasingly demands for efficient training methods. This paper describes an experiment exploring the use of haptic interaction for the purpose of skill training (haptic training). Various experiment tasks that include simple and complex tool paths have been developed for this purpose. 105 acquisition sessions distributed in 7 different tasks from 27 naive subjects and one surgeon performed teleoperated exercises with Omni Phantom. Task’s learning curves with and without robotic assistance and effects of damping have been discussed. For the force guidance case, assistance level gradually decreased the applied force as the training progressed. In the controversial scenario of robotic assistance for motor-learning benefits, this study shows such assistance improve the rate of learning for both simple and complex tasks.

Salih Ertug Ovur, Marisa Cobanaj, Luca Vantadori, Elena De Momi, Giancarlo Ferrigno
Augmented Reality Toolkit for a Smart Robot-Assisted MIS Platform

Minimally invasive surgery (MIS) offers important benefits to the patient, however it introduces additional complexity and constraints to the surgical workflow. Thus, the development of guidance systems which use Augmented Reality (AR) in order to provide clinicians with important information, has become a popular research topic in recent years. In this work, a visualization toolkit is presented, in which pre-operative 3D structures are overlaid to the intra-operative surgical view, during Robot-Assisted MIS. These structures are extracted from pre-operative Magnetic Resonance Imaging (MRI) or Computed Tomography (CT) scans and are registered into the endoscopic view through a semi-automatic pipeline, that includes minimal user interaction.

Georgios Zampokas, Konstantinos Tsiolis, Georgia Peleka, Angeliki Topalidou-Kyniazopoulou, Ioannis Mariolis, Sotiris Malasiotis, Dimitrios Tzovaras
Control of a da Vinci EndoWrist Surgical Instrument Using a Novel Master Controller

A novel master controller for robot-assisted minimally invasive surgery (RAMIS) is introduced and used to control a da Vinci EndoWrist instrument. The geometric model of the master mechanism and its mapping to the geometry of the EndoWrist tool are derived. Experimental results are conducted to open and close the jaws of an EndoWrist tool, and show that the developed mapping algorithm is accurate with a root mean square error of 0.7463 mm.

Sajeeva Abeywardena, Efi Psomopoulou, Mohammad Fattahi Sani, Antonia Tzemanaki, Sanja Dogramadzi
Toward a Neural-Symbolic Framework for Automated Workflow Analysis in Surgery

Learning production rules from continuous data streams, e.g. surgical videos, is a challenging problem. To learn production rules, we present a novel framework consisting of deep learning models and inductive logic programming (ILP) system for learning surgical workflow entities that are needed in subsequent surgical tasks, e.g. “what kind of instruments will be needed in the next step?” As a prototypical scenario, we analyzed the Robot-Assisted Partial Nephrectomy (RAPN) workflow. To verify our framework, first consistent and complete rules were learnt from the video annotations which can classify RAPN surgical workflow and temporal sequence at high-granularity e.g. steps. After we found that RAPN workflow is hierarchical, we used combination of learned predicates, presenting workflow hierarchy, to predict the information on the next step followed by a classification of step sequences with deep learning models. The predicted rules on the RAPN workflow was verified by an expert urologist and conforms with the standard workflow of RAPN.

Hirenkumar Nakawala, Elena De Momi, Roberto Bianchi, Michele Catellani, Ottavio De Cobelli, Pierre Jannin, Giancarlo Ferrigno, Paolo Fiorini
Manipulation of a Whole Surgical Tool Within Safe Regions Utilizing Barrier Artificial Potentials

Active constraint enforcement in robotic-assisted surgery is critical for reducing the intra-operative risk of unintentionally damaging sensitive tissues by the surgical instrument. This work considers surgical instruments which can be circumscribed by a geometric capsule and forbidden regions which can be approximated by point clouds in order to produce a repulsive wrench by the control action to guarantee manipulation within safe regions. This work details the control scheme which is based on barrier artificial potentials when considering the whole tool extending our previous results on the tool point. A proof of the control system’s passivity and non constraint violation is provided together with experimental results using a 7-dof KUKA LWR4+ manipulator as a master device in a virtual surgical scene in order to demonstrate the effectiveness of the proposed scheme.

Theodora Kastritsi, Iason Sarantopoulos, Sotiris Stavridis, Dimitrios Papageorgiou, Zoe Doulgeri
Evaluation of Force Feedback for Palpation and Application of Active Constraints on a Teleoperated System

A desktop haptic device is used to teleoperate an industrial redundant and compliant robotic arm with a surgical instrument mounted on its end-effector. The master and slave devices are coupled in a bilateral position-position architecture. Force feedback is provided by the master haptic device to the user, from the position of the slave’s wrist. A surgical task (palpation) that involves force feedback is presented and tested in a user study with surgeons and non-medical participants. Results show that users easily discern between three different materials during palpation given minimal familiarisation time. Active constraint enforcement is also integrated with the system as a sensitive area around the palpation samples which the slave instrument is prohibited to enter.

Efi Psomopoulou, Raj Persad, Anthony Koupparis, Sajeeva Abeywardena, Mohammad Fattahi Sani, Chris Melhuish, Sanja Dogramadzi
A Knowledge-Based Graphical Interface for Modeling Surgical Workflows in Robot-Assisted Minimally Invasive Surgery

This paper presents a knowledge-based graphical interface for modelling surgical workflows in robot-assisted minimally invasive surgery (R-A MIS). The interface provides an effortless way for the surgeon to generate a preoperative protocol for the surgery, while the knowledge base adds semantic information and rules to the surgery-specific ontologies. The surgeon can either modify a pre-existing prototype workflow for a known type of R-A MIS surgery or create a completely new surgical workflow for unknown types of R-A MIS. The workflow consists of phases and steps in consecutive order, incorporating anatomies, instruments and actions that the surgeon can modify. Any modification adheres to the rules of the knowledge base where information of incorporated robotic interactions is included. The interface is built as a web application that can be operated through a computer or portable device. The interface had positive reviews by orthopedic surgeons that were requested to create surgical workflows for R-A MIS arthroscopic procedures and evaluate its usability.

Christos Papadopoulos, Angeliki Topalidou-Kyniazopoulou, Ioannis Mariolis, Aristotelis Sideridis, Emmanouel Papacostas, Dimitrios Tzovaras
Augmented and Virtual Reality in Minimally Invasive Surgery, State of the Art and Future Prospects

Computer assisted guidance systems which utilize Augmented Reality (AR) and/or Virtual Reality (VR) during Minimally invasive surgery (MIS) have turned into a prominent research subject in the recent years. These technologies provide noteworthy advantages to the patients and the clinicians, by providing significant additional data during MIS. In this way, a priori information known to the surgeon pre-operatively (segmented anatomical structures of interest and lesions) can be superimposed semi-transparently on the directly visible tissues. This process can lead to higher preservation rates of healthy tissue and shorter post surgical recovery periods. The paper at hand is a survey investigating research efforts that have been conducted in the recent years in terms of utilization of AR and VR in laparoscopic and arthroscopic MIS.

Michele Catellani, Giovanni Cordima, Ottavio de Cobelli, Efthymios Papasoulis, Emmanuel Papacostas, Aristotelis Sideridis, Georgia Peleka, Georgios Zampokas, Konstantinos Tsiolis, Angeliki Topalidou-Kyniazopoulou, Ioannis Mariolis, Sotiris Malasiotis, Dimitrios Tzovaras

Special Sessions: Intelligent Computational Systems in Biomedical Engineering

Modeling and Objectification of Skiagraphy Image Quality Deterioration Caused by X-Ray Secondary Irradiation on Mobile X-Ray Device

Nowadays, the use of X-ray is an important part of imaging techniques used in medicine. The resulting skiagraphy images are kept in the workplace and can be influenced by secondary radiation, which occurs as a secondary effect when irradiated with X-rays. This secondary radiance has not to effect only for the skiagraphy images but also for the human body. The effect of secondary radiation is irreversible image acquisition, which can lead to poor diagnosis and poor clinical information. Currently, there are no techniques and methods available to determine the extent of secondary damage to the skiagraphy cassette. In the work, we analyse the influence of secondary radiation on the quality of the skiagraphy images using mobile X-ray devices. During image analysis, a multi-region segmentation method is used to determine pixels that are affected by irradiation. By this method, we are able to derive the characteristics of the image, determine the level of irradiation and estimate the effect of secondary radiation.

Klara Fiedorova, Martin Augustynek, Jan Kubicek, Marek Penhaker, Andrea Vodakova, Karol Korhelik
Segmentation of Blood Vessels from Fundus Retinal Images by Using Gabor Transformation

In this paper, it was using segmentation based on Gabor transformation for extraction the retinal blood vessels. The algorithm has been applied on images from freely accessible DRIVE and STARE databases. The segmentation algorithm provides the possibility of automatic modeling and consequent extraction of the retinal vascular system. Resultant images are in the form of a binary map. An algorithm for calculating the value of the curvature of the retinal vessels and thus objectifying the tortuosity can be applied to these images.

Alice Krestanova, Jan Kubicek, Jana Kosturikova
Evaluation of System for Simultaneous Measurement of Physiological Parameters: Potential for Determination of Age-Related Cardiovascular Status

We have developed a system for simultaneously determining arterial parameters by performing measurements on just one finger. The purpose of the present study was to evaluate the ability of this system to identify a state of ill health. The system can obtain eight physiological parameters: the systolic blood pressure (SBP), mean blood pressure (MBP), diastolic blood pressure (DBP), pulse pressure (PP), pulse rate (PR), normalized pulse volume (NPV) as an index of sympathetic activity, along with volume elastic modulus (Ev) and vascular volume change ratio (ηv). In this study, we attempted to identify poor physical condition and perform a clinical evaluation using these two parameters in particular. Poor physical condition includes the general deterioration in health due to age. We investigated whether it was possible to determine differences in health between old and young people. All participants were women because we wanted to eliminate errors due to gender, and measurements were performed under the same conditions. The results indicated that there were differences in the measured SBP, PP, Ev and ηv values for the older and younger groups, suggesting that the proposed system can be used to determine the health condition of a subject.

Honoka Koga, Jihyoung Lee, Peter Rolfe, Ken-ichi Yamakosh, Akira Kamiya, Takehiro Yamakoshi

Special Sessions: INT4DAT - Intelligent Systems and Technologies for Diagnostic, Assistance, and Therapeutics

Expressive Robotic Head for Human-Robot Interaction Studies

In this paper we present an improvement of the ISR-RobotHead prototype interaction capabilities having in view future Human-Robot Interactions studies. To accomplish that, a new hardware and software architectures and new facial expressions were developed. This new prototype (ISR-RobotHead v2.0) uses LCDs to display six human facial emotional expressions, avoiding the use of interlinked mechanical systems. The prototype also incorporates cameras, microphones, speakers, and LEDs on the robot’s cheeks. To validate the new cartoon facial expressions, an exploratory study was performed with children (4–6 years old) and adults (23–24 years old), who were asked to identify the expression displayed by the ISR-RobotHead v2.0; for children, the evaluation of accuracy in the detection of emotional expressions was performed through a corresponding number of cards picturing child’s emotional facial expressions. Preliminary results successfully validate the proposed robot’s non-verbal communication.

Ricardo Pereira, Luís Garrote, Tiago Barros, Carlos Carona, Luís C. Bento, Urbano J. Nunes
Machine Support to Discrimination of Parkinson’s Disease and Essential Tremor

Pathological tremor is a common but also highly complex movement disorder, affecting about 5% of population over the age of 65. Different methodologies have been proposed for its quantification and analysis. Nevertheless, the discrimination between Parkinsonian and Essential tremor remains a looming challenge in clinical practice, greatly impacting both patient treatment and the development of new interventions. In this study, the discriminative powers of cortical thickness features, extracted from magnetic resonance imaging of 21 ET, 15 PD patients and 18 healthy control subjects have been presented and discussed. A total of 129 volumetric features (whole brain, except cerebellum) and 152 cortical thickness features (average plus standard deviation of the thickness of the different cortical areas) have been extracted from each subject. Data mining of these characteristics has been employed to identify the most discriminative structural features for tremor diagnosis. These features were then combined in advanced classification models that yielded accuracy of up to 100% in discrimination of examined ET and PD patients, with the volume and thickness of enthorinal cortex as the most discriminative feature. Noteworthy, when compared to healthy age-matched controls, the identified MRI-based features indicate the structural changes of the brain and cortex in both PD and ET group of patients, potentially illuminating the neurodegenerative nature of both pathologies studied. This is an important finding by itself as the neurodegenerative aspects of ET have not yet been rigorously proven.

José Ignacio Serrano, Julián Benito-León, Aleš Holobar, Eduardo Rocon
Investigating Whole-Brain MRI Markers in Multiple Sclerosis – Emerging Dimensions in Morphometric Space

Magnetic resonance imaging (MRI) has become a crucial criterion for diagnosis and prognosis in early Multiple sclerosis (MS). Grey matter (GM) pathology has been described. However, little is known about the cause of cortical atrophy in MS. This study aims at identifying structural neuroimaging biomarkers of interest for the investigation of MS pathophysiology. We used structural MRI-based features - GM and WM volume, cortical thickness and gyrification index - to investigate patterns of deficits in MS. We analyzed images from 59 MS patients and 64 age-matched healthy controls. Imaging data underwent univariate statistical analyses, namely voxel-based morphometry (VBM) and surface-based morphometry (SBM), to investigate regional morphometric differences. Multivariate pattern analysis (MVPA) using a Support Vector Machine (SVM) classifier was also applied to explore pattern recognition in neuroimaging as a tool with potential for development of medical imaging biomarkers. Either VBM and SBM analyses yielded several morphological disease-related changes, which can be used to highlight disease effects in MS. SVM classification yielded an accuracy of 86,51% (sensitivity 74,58%, specificity 98,44%) with GM volume, while with WM data, SVM correctly classified 82,34% of participants (sensitivity 67,80%, specificity 96,88%). As such, MVPA can be a useful predictive biomarker with potential in assisting diagnosis in MS.

Júlia Soares, Teresa Sousa, Otília C. d’Almeida, Sónia Batista, Lívia Sousa, Miguel Castelo-Branco, João Valente Duarte
Computational Intelligence Generation of Subject-Specific Knee and Hip Healthy Joint Angles Reference Curves

Backpropagation Neural Network (BNN) and Extreme Learning Machine (ELM) can generate subject-specific joint angle reference profiles based on a subject’s height, weight, age and walking speed. These reference profiles are useful in various fields such as biomechanics and medicine, for detection of gait pathologies and rehabilitation. A common procedure used to identify abnormal gait is comparing an individual’s knee and hip curves against healthy reference curves. These reference curves are usually obtained from a heterogeneous sample of healthy subjects and might lack the specificity required to obtain accurate results. It is why the generation of reference curves according to an individual’s height, weight, age and gait speed shall result in a better comparison and diagnosis. The main objective of the present study is to observe which of the two Computational Intelligence (CI) methods, BNN and ELM, present more accurate results when used to generate reference curve profiles based on subject height, weight, age and gait speed for the knee and hip joint angles.

Pedro Sá Cunha, João Ferreira, A. Paulo Coimbra, Manuel Crisóstomo
Assistive Smart Cane (ASCane) for Fall Detection: First Advances

The development of fall detection systems with the capability of real-time monitoring is necessary considering that a large amount of people die and suffer severe consequences from falls. Due to their advantages, daily life accessories can be a solution to embed fall-related systems, and canes are no exception. In this paper, it is presented a cane with fall detection abilities. The ASCane is instrumented with an inertial sensor which data will be tested with three different fixed multi-threshold fall detection algorithms, one dynamic multi-threshold and machine learning methods from the literature. They were tested and modified to account the use of a cane. The best performance resulted in a sensitivity and specificity of 96.90% and 98.98%, respectively.

Pedro Mouta, Nuno Ferrete Ribeiro, Cristina P. Santos, Rui Moreira
Multi-view Robust Gesture Recognition for Assistive Interfaces

In this paper, we propose a gesture recognition approach using a multi-view setup for assistive device applications. As smart assistances become a reality, the need to interact with them in a natural fashion, as we do with other humans, becomes crucial. Gestures are a fundamental modality of human interaction, being natural and intuitive. We propose a gesture recognition approach relying on upper-body joints’ motions, so that individuals suffering from motor dysfunctions, that need to use wheelchairs or cannot stand, can as well interact with their smart assistive devices. To achieve this goal, we propose a robust multi-view skeleton fusion through a Kalman filtering technique, followed by an upper-body handcrafted feature extraction process. Gestures are classified using a support vector machine (SVM) classifier. Experiments with our captured dataset revealed a strong generalization from our method, and an increased performance of our multi-view fusion over the individual views.

João Paulo, Pedro Girão, Paulo Peixoto
Virtual Interface for an Active Motorized Pedal Exerciser for Human Leg Rehabilitation

The ability to move the lower limbs is crucial for most daily activities so there is a permanent need to develop new methods to improve people’s mobility. Common leg rehabilitation devices are pedal exercisers and static bicycles. This paper presents a tool for legs rehabilitation that is based on a motor assisted static bicycle, which is gradually triggered according to the pressure exerted on force sensors located in the pedals. It allows compensation for a leg with mobility limitations permitting the expected cycling movement to be performed easily. This device has also a sensor to monitor the patient’s heart rate. It is used to ensure the treatment’s efficiency and the patient’s safety. Each leg can be trained individually using different parameters. This is particularly useful for stroke patients. If the patient has a leg with mobility problems, the system compensates that leg impairment.The physiotherapist can manage the training parameters (speed, force on each pedal, heart rate) on a computer interface and monitor the training session. There were developed two interfaces for the patient to keep him motivated to stay close to the targeted values. The first interface consists of bar graphs that represent the values of feet forces, speed and heart rate, in real time. The second interface is a virtual game, a bicycle on a road, symbolizing the performance of the user as it would be in real life. The farther the patient’s performance is to the reference values, the fastest he or she will lose points. The game goal is to have the bicycle upright and the road with no inclination. Through ergonomic questionnaires it was determined that the virtual game was the preferred interface.

João Ferreira, A. Paulo Coimbra, Manuel Crisóstomo, Tao Liu

Special Sessions: Upper Limb Exoskeletons for a Better Quality of Life: What is Real, What is Useful, and What is Next?

Research Technologies for Assistance During Daily Life Activities

An overview of upper limbs exoskeletons aimed at supporting daily life activities of persons with different levels of disabilities is presented, with a particular focus on the functional gain the devices possibly deliver to the user. Observed differences in terms of effective functional improvement and self-perceived functional improvement will be discussed, with particular attention to deriving requirements for a successful assistive device.

Marta Gandolla, Alberto Antonietti, Valeria Longatelli, Stefano Dalla Gasperina, Emilia Ambrosini, Alessandra Pedrocchi
Clinical Needs and Possible Perspectives in Rehabilitation Context

One of the major fields of application of upper limb exoskeletons in rehabilitation medicine is to reduce disability enhancing recovery of functions in patients with Upper Motor Neuron Syndrome. Robotic treatment has received significant attention because it can provide high-intensity and repetitive movement therapy. There is much evidence that these technologies do not just provide assistance to perform functional exercises but are also involved in promoting re-learning processes by stimulating neuroplasticity. Human-machine interactions, control strategies, driving modes, mechanical design and training modes are scientific and technological challenges open in the community to improve the efficacy of robotics in neurorehabilitation to restore ability to manage daily life activities.

Franco Molteni, Roberto Ballarati, Eleonora Guanziroli
Upper-Limb Exoskeletons for Stroke Rehabilitation

Upper-limb exoskeletons provide high-intensity, repetitive, task-specific, interactive and individualized training, making effective use of neuroplasticity for functional recovery in neurological patients. Most exoskeletons have robot axes aligned with the anatomical axes of the subject and provide direct control of individual joints. Recently, novel mechanical structures and actuation mechanisms have been proposed, but still result in bulky and heavy exoskeletons, limiting their applicability into clinical practice. Technological efforts are needed to promote light and wearable exoskeletons that implement active-assistive controllers, providing “assisted-as-needed” rehabilitation therapy, towards patient’s motivation and self-esteem. An overview of upper-limb exoskeletons, including mechanical design and control algorithms, will be provided. Special focus will be put on the current evidence about the efficacy of wearable robotic technologies on motor recovery and about other therapies that can be combined with exoskeletons to improve their therapeutic effects.

Emilia Ambrosini, Stefano Dalla Gasperina, Marta Gandolla, Alessandra Pedrocchi
Industrial Wearable Robots: A HUMANufacturing Approach

In the last decades, exoskeletons have mostly been developed and studied for applications in the medical field, as rehabilitation or assistive devices for patients with movement disabilities. Recently, given the high performance of emerging wearable technologies, new applications have been proposed including the every-day support of able-bodied subjects such as workers. The execution of repetitive operations or actions that require excessive effort are the main causes of musculoskeletal injuries in people working in production lines or construction sites. The Industry 4.0 program is bringing companies to re-think their processes by considering human factors, ergonomics and sustainability issues. This is leading to a new trend in automation, which place the workers at the center of a modern smart factory, allowing them to take advantages of new interconnected tools. This new tendency fully embodied Comau’s vision. The company, in fact, has coined a term for better picturing its vision: HUMANufacturing. In this framework exoskeletons have the potential to become more and more adopted by industries as tools to provide support to the workers, preventing the rising of musculoskeletal diseases. This paper provides an overview of the main drivers of this nascent technology. Specifically, it aims to define the requirements that led to the development of an industrial exoskeleton, considering both the end-users and the manufacturers perspective, and showing how the HUMANufactuing approach has a role during the development of a new product.

Gaia Salvadore, Edoardo Rota, Elena Corsi, Giuseppe Colombina
Upper Limb Exoskeletons for a Better Quality of Life: What Is Currently Available, and What Is Missing in the Market

The market of medical devices for assisting disabled people at home and for rehabilitation in hospital and/or at home is very active and promising. The marketing approach, so far, has been based on experience-based report but a strong need to move to an evidence-based approach is one of the major changes that is expected in the next years. In addition, regulations are pushing towards this direction. How can we get ready for this challenge?

Marta Baratto, Claudio Ceresi, Valeria Longatelli

Special Sessions: Neurosystems and Connectivity

Optimization of a Motor Imagery Paradigm for Self-modulation of Bilateral Premotor Interhemispheric Functional Connectivity in fMRI Neurofeedback

Neuromodulation of a single brain region in an fMRI Neurofeedback experiment is a widely used technique with reported training effects on neuroplasticity and in behavioral changes. Recently, experiments aiming for modulation of functional relationship between segregated brain areas have been tested for their feasibility to understand underlying connectivity-based processing brain mechanisms. We performed a series of pilot studies to optimize an experimental protocol for interhemispheric functional connectivity modulation between bilateral premotor areas in a fMRI Neurofeedback setup, using a motor imagery paradigm. We have found that a better differentiation of correlation coefficient distributions between conditions was achieved with an optimized strategy with a reference task that consisted of the imagination of hand movement with gradual variation in the frequency of the imagined movements. Results also show that distinction between bimanual and unimanual motor imagery is not ideal for correlation-based differentiation using this paradigm. With this proof-of-concept study we conclude the feasibility of this setup as well as hypothesize that in future acquisitions, the reference tasks for an up-regulation condition should be based on this adaptive strategy and down-regulation condition should be optimized to promote the absence of activation in premotor areas.

João Pereira, Bruno Direito, Alexandre Sayal, Carlos Ferreira, Miguel Castelo-Branco

Special Sessions: Therapeutic Applications of Imaging and Neuro-Stimulation

A Hybrid Brain-Computer Interface Fusing P300 ERP and Electrooculography

An Electrooculography-based method is used to correct misclassification of P300 event related potentials in a Lateral Character Speller (LSC) Brain-Computer Interface (BCI). The LSC speller’s circular layout allows us to combine P300 detection with the detection of eye movements to improve symbol detection reliability. We separately classify the vertical and horizontal components of Electrooculography signals from shifts in user gaze during intertrial intervals, determining the quadrant of the character the participant will focus on in the next trial. A P300 EEG-based classification decision can then be corrected using quadrant information, selecting the character with the highest probability on that quadrant. This paper focuses on the implementation of the EOG quadrant detector. Preliminary results show good lateral identification but a lower selection accuracy. Empirically, it was possible to conclude that a relatively high percentage of P300 classification errors were corrected using lateral information alone, significantly increasing LSC character selection accuracy.

João Perdiz, Aniana Cruz, Urbano J. Nunes, Gabriel Pires
Non-invasive Spinal Cord Stimulation: Relevance of Modelling Studies in Clinical Protocol Design

Over the past two decades a growing interest has arisen in non-invasive neuromodulation of the spinal cord. The application of transcutaneous electric currents is a possible alternative therapy to many invasive procedures. This neuromodulatory technique may also be applied jointly with other therapies, such as pharmacologic substances, cognitive training and physiotherapy, to enhance the outcomes in the treatment of several neural dysfunctions. Electric stimulation originates electric fields in neurons that can change the transmembrane potential, facilitating or inhibiting neural responses. Varying stimulation parameters, geometry and placement of the stimulation devices (e.g. electrodes, coils) will result in different electric field patterns and, consequently, in different spinal neuron responses. We will address how computational modelling methods using realistic human models can be useful to optimize stimulation parameters for specific clinical purposes.

Sofia Rita Fernandes, Mariana Pereira, Mamede de Carvalho, Pedro Cavaleiro Miranda

Special Sessions: Value-Based Health Technology Assessment

Integrating HTA Principles into Procurement of Medical Devices: The Italian National HTA Programme for Medical Devices

Managing the adoption and diffusion of technological innovation is a key challenge for healthcare systems. In 2015, Italy has started designing and implementing a National Health Technology Assessment (HTA) Programme for Medical Devices (MDs), a network aimed at promoting the use of HTA tools, and safety, effectiveness and cost-effectiveness principles in the decision-making process of medical technologies. Since 2017, all the relevant stakeholders involved in HTA (i.e., patients and citizens’ representatives, scientific organizations, industry, healthcare organizations, academic researchers) are supporting a Steering Committee active at the Ministry of Health in the design and implementation of the Programme. Several working groups were established, one of them - based on a review of best practices - was instructed (i) to develop methods and procedures to integrate the results of HTA into decisions of procurement and clinical pathways and (ii) to propose a purchase request form template to be used by health care professionals. The final goal of the HTA Program is to reduce the heterogeneity across practices which translate into uneven access to innovative health technologies. This work provides an overview of the rationale and participatory process undertaken in Italy, and illustrates the working group main findings and recommendations.

Giuditta Callea, Carlo Federici, Oriana Ciani, Fabio Amatucci, Ludovica Borsoi, Rosanna Tarricone, Marcella Marletta
Multiple Criteria Decision Analysis for Health Technology Assessment of Medical Devices: A Winning Hospital-Based Experience

Health technology assessment (HTA) refers to the systematic evaluation of properties, effects, and/or impacts of health technologies. It is a multidisciplinary process to evaluate the social, economic, organizational and ethical issues of a health intervention or health technology, aiming to inform a policy decision making and to support decisions pertaining to the allocation of resources. In order to empower decision makers to choose more knowingly between the different alternatives, offering a more precise and more structured output as well as contextualized evidence for a specific technology, we developed a standardized methodological approach that integrates Multicriteria Decision Analysis (MCDA) with the EuNetHTA Core Model®. This new assessment approach, compared to the Core Model®, supplies a timelier as well as contextualized evidence for a specific technology. It makes it possible to obtain data which are more relevant and easier to interpret, and therefore more useful for decision makers to make investment choices with greater awareness. The application of MCDA to HTA has been mainly devised to address decision-making issues at hospital levels as the numerical values of the technologies’ performances cannot be always easily predictable and/or measurable being affected by the robustness of the evidences, which supports them. To overcome these issues, the proposed method, combines the static MCDA method with a dynamic Montecalo simulation allowing a structured and more precise output, giving the decision makers the possibility to knowledgeably choose between the different alternatives considered, often in a very short time.

Martina Andellini, Roxana di Mauro, Francesco Faggiano, Pietro Derrico, Matteo Ritrovato
Biodegradation Behavior of Magnesium Alloy During Exposure to the Conditions of Human Body Environment

In this study, biodegradation behaviour of WE43 magnesium alloy have been studied and compared during exposure to three different media commonly used to simulate the conditions of human body environment. Magnesium alloys emerged as a new class of bioresorbable implant materials. Their applications reduce certain risks associated with conventional permanent implants. Biodegradation behaviour of the WE43 magnesium alloy was observed under Dulbecco Modified Eagle Medium (DMEM, Sigma Aldrich) supplemented with 5% fetal bovine serum and gentamicin antibiotic as standard. The samples were stored in the medium at 37 °C and in a 5% CO2 atmosphere. The second type of medium was Hank’s Salt Balanced Solution (HBSS, Sigma Aldrich), which simulates the inorganic composition of blood plasma. HBSS was tempered at 37 °C. The last solution was an acidic solution of HCl + NaCl (pH2) with pH ~ 2 (0.01 M HCl and 0.14 M NaCl) tempered at 37 °C. The acidic type of solution was used to simulate a local acidic environment associated with osteoclast activity during bone remodelling. Changes in the mechanical properties of the samples during exposure to simulated body conditions were observed.

Radek Sedlacek, Tomas Suchy, Zdenek Padovec

Special Sessions: International Collaborative on Medical Devices Assessment

HTAi’s Role in the International Collaborative on Medical Device Assessments

HTAi – Health Technology Assessment international, was established in 2003 with the main objective to advance and promote HTA. The Society provides an open platform for global collaboration to leverage the shared and collective experience of its members to improve health outcomes globally. HTAi represents 82 organizational members, and 2,500 individual members from 65 countries. Members include leading HTA professionals and stakeholders, including researchers, policy makers, industry, academia, health service providers, agencies, and patients. The Society’s flagship event is an Annual Meeting that provides a platform for the global HTA community to come together to elevate global health decision-making in an open and collaborative environment. HTAi is also home to the Global Policy Forum, which offers an opportunity for senior-level individuals from public and private sector organizations using HTA to support decisions or recommendations about product development and coverage to interact with one another, HTAi Board members, and invited international experts. As well, HTAi has established 10 Interest Groups (IGs) to share international experiences and expertise among HTA users and producers globally. As a non-state actor in official relations with the WHO, HTAi is uniquely positioned to bring insight and expertise to further the work of the ICMDA. The newly established Medical Devices Interest Group and the HTA in Developing Countries IG also offers a platform within HTAi for discussion of the HTA needs, challenges, and solutions pertaining medical device use and HTA in developing countries.

Rebecca Trowman, Julie Polisena

Special Sessions: Ocular Imaging

Towards Improving Human Corneal Care Using Two-Photon Imaging

The human cornea is the tissue of the eye that contributes most to its refractive power. A healthy tissue is highly important to maintain visual acuity. In fact, diseases affecting the cornea are one of the major causes of blindness. Therefore, efficient methods to analyze the status of the healthy and diseased cornea are of outmost importance. Current clinical devices are, however, mostly limited to the morphological analysis of the tissue. The simultaneous analysis of tissue morphology, metabolism, and stromal structural organization using two-photon imaging (TPI) could improve corneal examination. In this study, two systems with TPI capabilities equipped with ultra-short near-infrared Ti:sapphire lasers were optimized for corneal imaging and their advantages for tissue examination demonstrated. Additionally, we show that by using TPI, disease diagnosis, follow-up after medical procedures, and corneal evaluation prior to transplantation could be improved. Thus, a future clinical device based on TPI could enhance the current state corneal examination and improve patient diagnosis and care.

Ana Batista, Hans Georg Breunig, Berthold Seitz, Karsten König
Characterization of the Retinal Changes of the 3xTg-AD Mouse Model of Alzheimer’s Disease

In this work, we imaged the retina of wild-type and triple-transgenic mice model of Alzheimer’s disease (AD) (3xTg-AD), at the ages of one and two months, by optical coherence tomography. Texture analysis of calculated fundus images, for the six most anterior layers of the retina, present widespread differences between groups, demonstrating that retinal changes in the animal model of AD are notorious across the neuroretina. These results and the consistency of data pave the way for the identification of texture biomarkers of disease onset and progression with potential application to the human retina.

Hugo Ferreira, João Martins, Ana Nunes, Paula I. Moreira, Miguel Castelo-Branco, António Francisco Ambrósio, Pedro Serranho, Rui Bernardes
Distinguishing Functional from Non-functional Pituitary Macroadenomas with a Machine Learning Analysis

Pituitary adenomas are rare intracranial tumors that are often found incidentally in MR images. On the other hand, radiomics is a new field whose aim is converting images in mineable data; particularly, texture analysis is a postprocessing technique extracting quantitative parameters from the heterogeneity of pixel grey level. In this scenario, machine learning can be applied in order to classify these adenomas into functional and non-functional starting from features extracted through texture analysis on MRI images acquired through a protocol including a coronal T2-weighted Turbo Spin Echo sequence. The boosting of J48, a multinomial logistic regression and K nearest neighbour are implemented employing Knime analytics platform. Excluding J48 whose accuracy was 83.0%, multinomial logistic regression and K nearest neighbour achieved accuracies beyond 92.0% and the Area Under the Curve Receiving Characteristic Operator till 98.4%. Diagnosing correctly this delicate disease is crucial in order to achieve the best management as well as the most appropriate cure for patients. The novelty of this paper lies in proving the ability of the combination of radiomics and machine learning to pre-operatively predict tumoral behavior. Prior to this analysis it was believed that only blood tests or histopathological analysis could provide this information.

Ricciardi Carlo, Cuocolo Renato, Cesarelli Giuseppe, Ugga Lorenzo, Improta Giovanni, Solari Domenico, Romeo Valeria, Guadagno Elia, Cavallo Luigi Maria, Cesarelli Mario
Sexual Dimorphism of the Adult Human Retina Assessed by Optical Coherence Tomography

In this work, we imaged the retina of healthy age-matched female and male subjects with optical coherence tomography to assess the existence of localised sex-related differences in the retina. Texture analysis revealed statistically significant differences spread over the different layers of the neuroretina, particularly at the ganglion cell, inner plexiform and inner nuclear layers, and also spread in the macula. These results suggest that sexual dimorphism is expressed in the retina at the microstructural level.

Ana Nunes, Pedro Serranho, Hugo Quental, António Francisco Ambrósio, Miguel Castelo-Branco, Rui Bernardes

Challenges: IFMBE Scientific Challenge Competition

Convolutional Neural Network for a P300 Brain-Computer Interface to Improve Social Attention in Autistic Spectrum Disorder

A Brain-Computer Interface (BCI) relies on machine learning algorithms to decode the brain signals. An accurate detection of P300 response in electroencephalography (EEG) data can be used to design P300-based BCIs to improve social attention in Autistic Spectrum Disorder (ASD). Recently, there was a growing interest in the application of Convolutional Neural Networks (CNNs) to decode P300 in an end-to-end fashion. However, the complexity of these models needs to be carefully taken into account. In this study, a lightweight CNN previously validated for P300 detection (EEGNet) was used to decode which object ASD participants were paying attention to in a virtual environment. Two learning strategies were deepened: within-session and cross-session trainings. Cross-session training resulted in a higher target object accuracy scoring 92.27% on average across sessions and subjects, and in a lower decoding variability across sessions.

Davide Borra, Silvia Fantozzi, Elisa Magosso
Deep Learning Architecture Based on the Combination of Convolutional and Recurrent Layers for ERP-Based Brain-Computer Interfaces

In this study, we present a novel deep learning architecture for brain-computer interfaces based on event related potentials (ERP). The topology of the neural network combines convolutional and recurrent layers in order to learn high-level spatial and temporal features. Specifically, our model uses a convolutional layer, intended to detect spatial patterns over the scalp in short periods of time, followed by two bidirectional long-short term memory (BLSTM) layers to extract long-term temporal dependencies within the data. To the best of our knowledge, this is the first time that BLSTM layers are explored for ERP classification. This study takes part in the MEDICON 2019 IFMBE scientific challenge. The model has been evaluated using the provided dataset for the competition (15 subjects with autism spectrum disorder, 7 BCI sessions), achieving an average accuracy of 84% in command selection. In the course of our experiments, this approach outperformed traditional methods, such as step-wise linear discriminant analysis (SWLDA), and other deep learning architectures.

Eduardo Santamaría-Vázquez, Víctor Martínez-Cagigal, Javier Gomez-Pilar, Roberto Hornero
Slow Cortical Potential BCI Classification Using Sparse Variational Bayesian Logistic Regression with Automatic Relevance Determination

Detecting P300 slow-cortical ERPs poses a considerable challenge in signal processing due to the complex and non-stationary characteristics of a single-trial EEG signal. EEG-based neurofeedback training is a possible strategy to improve the social abilities in Autism-Spectrum Disorder (ASD) subjects. This paper presents a BCI P300 ERPs based protocol optimization used for the enhancement of joint-attention skills in ASD subjects, using a robust logistic regression with Automatic Relevance Determination based on full Variational Bayesian inference (VB-ARD). The performance of the proposed approach was investigated utilizing the IFMBE 2019 Scientific Challenge Competition dataset, which consisted of 15 ASD subjects who underwent a total of 7 BCI sessions spread over 4 months. The results showed that the proposed VB-ARD approach eliminates irrelevant channels and features effectively, producing a robust sparse model with 81.5 ± 12.0% accuracy in relatively short modeling computational time 19.3 ± 1.4 s, and it outperforms the standard regularized logistic regression in terms of accuracy and speed needed to produce the BCI model. This paper demonstrated the effectiveness of the probabilistic approach using Bayesian inference for the production of a robust BCI model. Considering the good classification accuracy over sessions and fast modeling time the proposed method could be a useful tool used for the BCI based protocol for the improvement of joint-attention ability in ASD subjects.

Aleksandar Miladinović, Miloš Ajčević, Piero Paolo Battaglini, Giulia Silveri, Gaia Ciacchi, Giulietta Morra, Joanna Jarmolowska, Agostino Accardo
A Feasible Classification Algorithm for Event-Related Potential (ERP) Based Brain-Computer-Interface (BCI) from IFMBE Scientific Challenge Dataset

Event-Related Potential (ERP) based Brain-Computer Interfaces (BCI) have been widely investigated as an alternative human computer interaction solution. For people with neurological diseases and severe disabilities like amyotrophic lateral sclerosis (ALS) and stroke, BCI may be their only access method for communication. For people with neurodevelopmental disorders, such as autism spectrum disorder (ASD), BCI is also considered to be a potential rehabilitation and education assistance method. Although these are promising developments, further work is required to optimize current classification and filtering methods to improve reliability and enhance the user experience. The aim of this project is to investigate the four most commonly used classification algorithms: Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), Convolutional Neural Network (CNN) and Long-Short Term Memory (LSTM) with a novel and personalised filter design. Results of the classification tasks in Phase I and Phase II, of the IFMBE Scientific Challenge, show a 73% accuracy for phase I and 67% accuracy for Phase II.

Haifeng Zhao, Shiduo Yu, Joseph Prinable, Alistair McEwan, Petra Karlsson
Linear vs Nonlinear Classification of Social Joint Attention in Autism Using VR P300-Based Brain Computer Interfaces

VR P300-based BCI has proven to be a suitable method for training social attention skills in youngsters with autism spectrum disorder (ASD). In this study, we present a method that could be used in such an application to identify which object the user is paying attention to in a virtual environment by means of EEG recordings only. Temporal and time-frequency features were explored. Furthermore, the prediction accuracy of linear and nonlinear classification methods was assessed and compared, along with their computational times and complexity, and linear discriminant analysis (LDA) yielded the best overall performance (82%). The successful predictions and low computational times demonstrate the feasibility of the proposed solution for a VR-BCI neurorehabilitation tool.

Lucia de Arancibia, Patricia Sánchez-González, Enrique J. Gómez, M. Elena Hernando, Ignacio Oropesa
Linear SVM Algorithm Optimization for an EEG-Based Brain-Computer Interface Used by High Functioning Autism Spectrum Disorder Participants

Machine-learning algorithms can be used for data classification on EEG-based Brain-computer interfaces (BCIs). Here, we used an algorithm based on linear support vector machine (SVM) to identify the presence of the P300 component in datasets from 15 young adult participants with autism spectrum disorder that were provided for the IFMBE Scientific Challenge 2019. We optimized the parameters and inputs for a linear SVM model throughout the ten attempts of the challenge and compared them in terms of accuracy. The highest score (mean accuracy) of 82% was achieved by a procedure that was customized per session per participant. When using a similar procedure for classification model generation and configuration of parameters for all sessions and participants, the highest score achieved was 77%. The results showed that adding data from targets from different calibration sessions from the same participants to the training dataset resulted in a significant increase in accuracy. In all attempts, the mean accuracy was above 70%, which is considered the minimum classification level for the controllability of a BCI. These are promising results for future use of BCIs as a tool for attention training in ASD participants.

Mayra Bittencourt-Villalpando, Natasha M. Maurits
Classification of P300 Component Using a Riemannian Ensemble Approach

We present a framework for P300 ERP classification on the 2019 IFMBE competition dataset using a combination of a Riemannian geometry and ensemble learning. Covariance matrices and ERP prototypes are extracted after the EEG is passed through a filter bank and an ensemble of LDA classifiers is trained on subsets of channels, trials, and frequencies. The model selects a final class based on maximum probability of evidence from all ensembles. Our pipeline achieves an average classification accuracy of 81.2% on the test set.

Dominik Krzemiński, Sebastian Michelmann, Matthias Treder, Lorena Santamaria
Using Time Domain and Pearson’s Correlation to Predict Attention Focus in Autistic Spectrum Disorder from EEG P300 Components

Patients with Autistic Spectrum Discorder are known to have deficits in interpreting others’ intentions from gaze-direction or other social attention. Here, we use electroencephalography data recorded in virtual reality experiments with patients to predict one out of eight objects that was focused on. Correct labels for these objects were known from parallel eye-tracking measurements. We extracted features from the time domain and from Pearson’s correlation and applied both statistical and neuro-inspired supervised machine learning algorithms. Using a multi-layer perceptron, we achieved 65.4% accuracy on the validation data set and 70.0% accuracy on the test data set.

V. Sophie Adama, Benjamin Schindler, Thomas Schmid
Performance Evaluation of Manifold Algorithms on a P300 Paradigm Based Online BCI Dataset

Healthcare field is highly benefited by incorporating BCI for detection and diagnosis of some health related detriment as well as rehabilitation and restoration of certain disabilities. An EEG dataset acquired from 15 high-functioning ASD patients, while they were undergoing a P300 experiment in a virtual reality platform, was analysed in this paper using three algorithms. Performance of Bayes Linear Discriminant Analysis (BLDA) was predominant over Convolutional Neural Network (CNN) and Random Undersampling (RUS) Boosting. BLDA rendered 73% overall accuracy in predicting target and the best accuracy for each subject using CNN or BLDA yielded an overall accuracy of 76%.

Bipra Chatterjee, Ramaswamy Palaniappan, Cota Navin Gupta

Challenges: Fraunhofer Best Portuguese PhD Thesis Competition in Biomedical Engineering

Magnetic Carbon Nanostructures and Study of Their Transport in Microfluidic Devices for Hyperthermia

Cancer incidence and mortality are growing worldwide at an alarming pace, emphasizing the urgent need for new strategies to combat this disease. One of the frontiers of cancer research is currently focused on the design of multifunctional magnetic nanoparticles capable to achieve the synergistic cancer theranostics (both diagnosis and therapy). Although the potentiality that these multifunctional nanosystems represents to nanomedicine, cancer treatment and diagnostic, there are still many challenges that must be addressed in a near future before this approach became a reality. The development of efficient multifunctional magnetic nanosystems able to selectively destroy cancer cells in detriment of healthy ones, is one of the main challenges that have damped the spread of this technology into clinical applications. The limited biological and biophysical studies between the biomedical nanosystems and cells/tissues/organs is another challenge that has to be addressed. With these two main challenges in mind, the present Ph.D. work was focused in the development of: (1) Multifunctional magnetic carbon nanostructures as multifunctional nanosystems for the treatment of cancer, and (2) New advanced microfluidic devices capable to give new insights over the developed nanosystems and human cells.

Raquel O. Rodrigues, Rui Lima, Helder T. Gomes, Adrián M. T. Silva
TESSEE – Tool for Early Stem Cells Economic Evaluation

Stem cell therapies are promising for diverse indications. However, there are manufacturing and reimbursement challenges that must be addressed toward widespread adoption. Relying on bioprocess and/or health economics models, early health technology assessment (eHTA) has provided insights on manufacturing innovations towards reducing the manufacturing cost of goods (CoG), and the long-term value that prospective stem cell therapies must provide to secure reimbursement. These economics models have focused on commercial tools or proprietary code. In order to increase awareness to the usefulness of eHTA in stem cell engineering, this thesis presents a new open-source tool, TESSEE (Tool for Early Stem Cells Economic Evaluation). Three relevant case studies for mesenchymal stem/stromal cell (MSC) and pluripotent stem cell (PSC) based therapies are presented as applications of TESSEE for decision-making in the implementation of stem cell engineering innovations. A study on the choice of a culture media supplement for autologous MSC therapy manufacturing determined that human platelet lysate (hPL) reduces CoG, in comparison to fetal bovine serum (FBS), for 97% of donors. An expansion system focused case study showed that the implementation of a new vertical wheel reactor (VWR) for the microcarrier-based culture of MSC increases the number of cells per batch and reduces costs of goods per dose by up to 48% from typical two-dimensional flasks for expansion. eHTA of devices containing PSC-derived beta cells for treatment of type 1 diabetes patients showed that the transplantation of the devices could be very effective at improving the quality of life. However, a price reduction, accompanied by reduction of manufacturing costs, is required to achieve widespread cost-effectiveness. This thesis highlights the versatility of TESSEE and builds evidence for eHTA for rational implementation of innovations in stem cell engineering, towards more cost-effective stem cell therapies.

Cátia Bandeiras, Joaquim Manuel Sampaio Cabral, Stan Neil Finkelstein, Frederico Castelo Ferreira

Challenges: Fraunhofer Best Portuguese MsC Thesis Competition in Biomedical Engineering

Community Finding with Applications on Phylogenetic Networks

With the advent of high-throughput sequencing methods, new ways of visualizing and analyzing increasingly amounts of data are needed. Although some software already exist, they do not scale well or require advanced skills to be useful in phylogenetics.The aim of this thesis was to implement three community finding algorithms – Louvain, Infomap and Layered Label Propagation (LLP); to benchmark them using two synthetic networks – Girvan-Newman (GN) and Lancichinetti-Fortunato-Radicchi (LFR); to test them in real networks, particularly, in one derived from a Staphylococcus aureus MLST dataset; to compare visualization frameworks – Cytoscape.js and D3.js, and, finally, to make it all available online ( ).Louvain, Infomap and LLP were implemented in JavaScript. Unless otherwise stated, next conclusions are valid for GN and LFR. In terms of speed, Louvain outperformed all others. Considering accuracy, in networks with well-defined communities, Louvain was the most accurate. For higher mixing, LLP was the best. Contrarily to weakly mixed, it is advantageous to increase the resolution parameter in highly mixed GN. In LFR, higher resolution decreases the accuracy of detection, independently of the mixing parameter. The increase of the average node degree enhanced partitioning accuracy and suggested detection by chance was minimized. It is computationally more intensive to generate GN with higher mixing or average degree, using the algorithm developed in the thesis or the LFR implementation. In S. aureus network, Louvain was the fastest and the most accurate in detecting the clusters of seven groups of strains directly evolved from the common ancestor.

Luís Rita, Alexandre Francisco, João Carriço, Vítor Borges
Functional Electrical Stimulation for Gait Rehabilitation

Conditions that can lead to a full or partial motor function loss, such as stroke or multiple sclerosis, leave people with disabilities that may interfere severely with lower body movements, such as gait. Drop Foot (DF) is a gait disorder that results in a reduced ability or total inability to contract the Tibialis Anterior (TA) muscle, causing an inability to raise the foot during gait. One of the most effective methods to correct DF is Functional Electrical Stimulation (FES). FES is a technique used to reproduce the activation patterns of functional muscles, in order to create muscular contractions through electrical stimulation of the muscle’s nervous tissue.FES has first been introduced in 1961. However, the available commercial FES systems still do not take into account the fact that the gait differs from subject to subject, depending on their physical condition, muscular fatigue and rehabilitation stage. Therefore, they are unable to provide a personalized assistance to the user, delivering constant stimulation pulses that are only based on gait events. Consequently, they promote the early onset of fatigue and generate coarse movements. This dissertation aims to tackle the aforementioned issues by developing a FES system for personalized DF correction, tailored to each individual user’s needs through the use of a Neural Network (NN). A Non-Linear Autoregressive Neural Network with Exogenous inputs (NARX Neural Network) was used to model the dynamics of the electrically stimulated TA muscle, in a novel approach that uses both the foot angle and the foot velocity. The model was combined with a Proportional Derivative controller to help compensate for any external disturbances. In order to create more natural movements, reference trajectories were obtained by recording the foot angle and velocity of healthy subjects walking at different speeds.The system has been validated with a healthy subject walking at 3 different speeds on a treadmill: 1 km/h, 1.5 km/h and 2 km/h. It was able to track the desired trajectory for every speed, thus creating a more natural movement and effectively correcting DF gait.

Ana Correia, Jorge M. Martins, Cristina P. Santos
Deep Aesthetic Assessment of Breast Cancer Surgery Outcomes

Breast cancer is a highly mutable and rapidly evolving disease, with a large worldwide incidence. Even though, it is estimated that approximately 90% of the cases are treatable and curable if detected on early staging and given the best treatment. Nowadays, with the existence of breast cancer routine screening habits, better clinical treatment plans and proper management of the disease, it is possible to treat most cancers with conservative approaches, also known as breast cancer conservative treatments (BCCT). With such a treatment methodology, it is possible to focus on the aesthetic results of the surgery and the patient’s Quality of Life, which may influence BCCT outcomes. In the past, this assessment would be done through subjective methods, where a panel of experts would be needed to perform the assessment; however, with the development of computer vision techniques, objective methods, such as BAT© and BCCT.core, which perform the assessment based on asymmetry measurements, have been used. On the other hand, they still require information given by the user and none of them has been considered the gold standard for this task. Recently, with the advent of deep learning techniques, algorithms capable of improving the performance of traditional methods on the detection of breast fiducial points (required for asymmetry measurements) have been proposed and showed promising results. There is still, however, a large margin for investigation on how to integrate such algorithms in a complete application, capable of performing an end-to-end classification of the BCCT outcomes. Taking this into account, this thesis shows a comparative study between deep convolutional networks for image segmentation and two different quality-driven keypoint detection architectures for the detection of the breast contour. One that uses a deep learning model that has learned to predict the quality (given by the mean squared error) of an array of keypoints, and, based on this quality, applies the backpropagation algorithm, with gradient descent, to improve them; another which uses a deep learning model which was trained with the quality as a regularization method and that used iterative refinement, in each training step, to improve the quality of the keypoints that were fed into the network. Although none of the methods surpasses the current state of the art, they present promising results for the creation of alternative methodologies to address other regression problems in which the learning of the quality metric may be easier. Following the current trend in the field of web development and with the objective of transferring BCCT.core to an online format, a prototype of a web application for the automatic keypoint detection was developed and is presented in this document. Currently, the user may upload an image and automatically detect and/or manipulate its keypoints. This prototype is completely scalable and can be upgraded with new functionalities according to the user’s needs .

Tiago Gonçalves, Wilson Silva, Jaime Cardoso
Deep Learning for Interictal Epileptiform Discharge Detection from Scalp EEG Recordings

Interictal Epileptiform Discharge (IED) detection in EEG signals is widely used in the diagnosis of epilepsy. Visual analysis of EEGs by experts remains the gold standard, outperforming current computer algorithms. Deep learning methods can be an automated way to perform this task. We trained a VGG network using 2-s EEG epochs from patients with focal and generalized epilepsy (39 and 40 patients, respectively, 1977 epochs total) and 53 normal controls (110770 epochs). Five-fold cross-validation was performed on the training set. Model performance was assessed on an independent set (734 IEDs from 20 patients with focal and generalized epilepsy and 23040 normal epochs from 14 controls). Network visualization techniques (filter visualization and occlusion) were applied. The VGG yielded an Area Under the ROC Curve (AUC) of 0.96 (95% Confidence Interval (CI) = 0.95 − 0.97). At 99% specificity, the sensitivity was 79% and only one sample was misclassified per two minutes of analyzed EEG. Filter visualization showed that filters from higher level layers display patches of activity indicative of IED detection. Occlusion showed that the model correctly identified IED shapes. We show that deep neural networks can reliably identify IEDs, which may lead to a fundamental shift in clinical EEG analysis.

Catarina Lourenço, Marleen C. Tjepkema-Cloostermans, Luís F. Teixeira, Michel J. A. M. van Putten
Feedback-Error Learning Control for Powered Assistive Devices

Active orthoses (AOs) are becoming relevant for user-oriented training in gait rehabilitation. This implies efficient responses of AO’s low-level controllers with short time modeling for medical applications. This thesis investigates, in an innovative way, the performance of Feedback-Error Learning (FEL) control to time-effectively adapt the AOs’ responses to user-oriented trajectories and changes in the dynamics due to the interaction with the user. FEL control comprises a feedback PID controller and a neural network feedforward controller to promptly learn the inverse dynamics of two AOs. It was carried out experiments with able-bodied subjects walking on a treadmill and considering external disturbances to user-AO interaction. Results showed that the FEL control effectively tracked the user-oriented trajectory with position errors between 5% to 7%, and with a mean delay lower than 25 ms. Compared to a single PID control, the FEL control decreased by 16.5% and 90.7% the position error and delay, respectively. Moreover, the feedforward controller was able to learn the inverse dynamics of the two AOs and adapt to variations in the user-oriented trajectories, such as speed and angular range, while the feedback controller compensated for random disturbances. FEL demonstrated to be an efficient low-level controller for controlling AOs during gait rehabilitation.

Pedro Nuno Fernandes, Joana Figueiredo, Juan C. Moreno, Cristina Peixoto Santos
CopyRobot: Interactive Mirroring Robotics Game for ASD Children

The family of disorders commonly known as autism is characterized by a deficit in social interaction and restricted repetitive and stereotyped patterns of behaviours, activities and interests. Motor disturbances are not part of the diagnosis of the children with autism but some studies have estimated that between 80 and 90% of children with Autism Spectrum Disorder (ASD) demonstrate some degree of motor impairments. Several therapies have been used for the improvement of motor skills, always leading to behavioural improvements as side-effects, demonstrating the importance of motor interaction and stimulation for the case of autism. Recent studies have shown that motor, imitation and social abilities are all related in people with autism. In this work, a humanoid robot is used to create a therapy that unites all these areas. The system involves a robot (NAO), a Kinect camera and Personal Computer, with the goal of facilitating the interaction between therapist and a child with ASD during a physical therapy session. To improve the imitation abilities of the child, the robot was programmed to mirror both the child and the therapist movements. After testing different tracking methodologies, the Kinect sensor was selected as the best compromise of quality and cost. Two protocols were developed, depending on who plays the role of the main actor. In the first protocol, the robot is the master and leads the interaction. It decides the exercise to execute and gives feedback to both the therapist and the child. In the second protocol, the choice of the exercise sequence is the therapist’s responsibility. To promote interaction further during clinical tests, the protocol was changed to include gesture imitation. For the robot master protocol, the space theme was chosen. For the therapist master protocol, the theme of sports, that was already performed by the children in the usual therapy, was adopted. The system was tested in realistic conditions with two different autistic children. The reaction was different in each case but it demonstrated the importance of these imitation games in the treatment of this disease.

Laura Santos, Alice Geminiani, Ivana Olivieri, José Santos-Victor, Alessandra Pedrocchi
Neuromechanical and Environment Aware Machine Learning Tool for Human Locomotion Intent Recognition

Current research suggests the emergent need to recognize and predict locomotion modes (LMs) and LM transitions to allow a natural and smooth response of lower limb active assistive devices such as prostheses and orthosis for daily life locomotion assistance. This Master dissertation proposes an automatic and user-independent recognition and prediction tool based on machine learning methods. Further, it seeks to determine the gait measures that yielded the best performance in recognizing and predicting several human daily performed LMs and respective LM transitions. The machine learning framework was established using a Gaussian support vector machine (SVM) and discriminative features estimated from three wearable sensors, namely, inertial, force and laser sensors. In addition, a neuro-biomechanical model was used to compute joint angles and muscle activations that were fused with the sensor-based features. Results showed that combining biomechanical features from the Xsens with environment-aware features from the laser sensor resulted in the best recognition and prediction of LM (MCC = 0.99 and MCC = 0.95) and LM transitions (MCC = 0.96 and MCC = 0.98). Moreover, the predicted LM transitions were determined with high prediction time since their detection happened one or more steps before the LM transition occurrence. The developed framework has potential to improve the assistance delivered by locomotion assistive devices to achieve a more natural and smooth motion assistance.

Simão Carvalho, Joana Figueiredo, Cristina P. Santos
Classification of Patients with Parkinson’s Disease Using Medical Imaging and Artificial Intelligence Algorithms

The diagnosis of Parkinsonian Syndromes (PS) at early-stages is a challenge. PS usually present similar symptoms, and the diagnosis is mainly clinical, causing often misdiagnosis between PS and other movement disorders. Parkinson’s Disease (PD) is the most common PS, affecting a large part of the worldwide population. Medical imaging such as Magnetic Resonance Imaging (MRI) and Single Photon Emission Computed Tomography (SPECT) are currently being used to detect changes in anatomy and in the dopaminergic system, respectively. SPECT imaging allowed to find a group of patients diagnosed as having PD but without the characteristic decreased uptake of a dopamine analogue, so called “Scans Without Evidence of Dopaminergic Deficit” (SWEDD). Nowadays, deep learning algorithms, such as Convolutional Neural Networks (CNN), are becoming a useful tool in the medical field to detect patterns relevant to diseases in images. This study proposed an approach using CNN for the classification of MRI and SPECT images from PD, SWEDD, and Control subjects, to identify regions-of-interest related to PD. The proposed model achieved an accuracy of 97.4% using MRI images encompassing the mesencephalon and 93.3% with SPECT slices encompassing the basal ganglia. The results suggest that CNN was able to discriminate Control vs. PD and PD vs. SWEDD, CNN achieved accuracies up to 65.7%. Regarding PD vs. SWEDD, this classification obtained an accuracy of 73.3% using MRI images encompassing the mesencephalon and 93.3% with SPECT slices embracing the basal ganglia. The results suggest that CNN was able to discrimination Control vs. PD and PD vs. SWEDD, but not Control vs. SWEDD supporting the fact that SWEDD patients do not show evidence of dopamine deficit. In addition, the classification allowed the identification of the images comprising the mesencephalon or the basal ganglia as the most relevant for the classification.

Helena R. Pereira, Hugo A. Ferreira
XV Mediterranean Conference on Medical and Biological Engineering and Computing – MEDICON 2019
Jorge Henriques
Nuno Neves
Prof. Dr. Paulo de Carvalho
Copyright Year
Electronic ISBN
Print ISBN