Computer Methods, Imaging and Visualization in Biomechanics and Biomedical Engineering II

In this paper a Spatial Markov Chain Cellular Automata model for the spread of viruses is proposed. The model is based on a graph with connected nodes, where the nodes represent individuals and the connections between the nodes denote the relations between humans. In this way, a graph is connected where the probability of infectious spread from person to person is determined by the intensity of interpersonal contact. Infectious transfer is determined by chance. The model is extended to incorporate various lockdown scenarios. Simulations with different lockdowns are provided. In addition, under logistic regression, the probability of death as a function of age and gender is estimated, as well as the duration of the disease given that an individual dies from it. The estimations have been done based on actual data of RIVM (from the Netherlands).

The temporomandibular joint (TMJ) is one of the most complex joints in the human body with its ability to rotate and translate during jaw opening and closing movements. The motion during jaw-closing has two phases: 1) rotation and translation along the mandibular fossa from maximum mouth open (MMO), and 2) pure rotation from 22° until mouth closed position. However, this motion pattern is altered after the jaw undergoes open joint arthroplasty. This study analyzes the motion and torque capabilities associated with the Fibula Free Flap (FFF) procedure, and two types of TMJ implants: 1) a generic implant constrained to pure rotation, and 2) a patient-specific (PS) implant capable of both rotation and translation. Results show the PS implant most resembles the healthy jaw showing a 34.4 mm inferior displacement at MMO; a 1.15% decrease compared to the healthy jaw’s (34.8 mm), and a maximum net torque of 18,600 N-mm; a 29.3% decrease compared to 26,300 N-mm. The rotation-constrained FFF procedure and generic implant limits the jaw to 70% of MC rotation decreasing the inferior displacement at MMO (FFF/Generic: 22.5 mm) by 35.3% compared to the healthy jaw and prohibits net torque about the MC after the pure rotation phase.

Research focusing on optimal rehabilitation methods has been directed towards powered lower-limb exoskeletons which combines the advantages delivered from the grounded robotic devices with the ability to train the patient in a real-world environment. In this context literature has highlighted the benefit of coupling gait analysis and musculoskeletal modeling for treatment planning. Recently, this combined approach has been successfully applied to detect the alterations in motor control related to Parkinson’s Disease (PD). However, no study has reported about the effects of an overground wearable exoskeleton in terms of both gait analysis and musculoskeletal modeling-derived parameters in people with PD. The aim of this study was to quantitatively assess the effect of an overground rehabilitation treatment on a PD participant both in terms of gait parameters and muscle forces. One people with PD has been enrolled and gait analysis was performed before and after a 4-weeks gait training intervention with an overground exoskeleton. Inverse kinematics, inverse dynamics, and static optimization were performed in OpenSim. Results from joint moments and muscle forces were compared with a group of healthy controls. Preliminary results showed that after the therapy joint loads in both ankle and knee joints were reduced during the stance phase and muscle forces displayed an increased magnitude in their peak after the treatment. For the best of our knowledge, the presented case study is the first attempt to track rehabilitation improvement via muscle forces assessment. Further studies should focus on increasing the sample size to generalize the outcomes.

Physical impairments have multiple causes, making them common. Locomotion disorders have been afflicting society for a long time, motivating researchers and engineers to mitigate their consequences. Nowadays, solutions such as exoskeletons and exosuits are in constant development and may become reliable options to help people in these circumstances. However, prospective solutions need a control system acting as a “bridge” between the external device (actuators) and the user. Among several possibilities, movement prediction is prioritized over movement reaction. This task may be done by capturing and processing biological signals from a user’s body. Within this paradigm, muscle electromyographic (EMG) signals were acquired, processed and sent as input to piezoelectric soft actuators.

In an effort to enhance medical decision-making and precision treatment, biomechanical musculoskeletal modeling and computer simulations originating from observations of human movements are being implemented. Such a combination allows for the establishment of cause-and-effect relationships providing a unique, non-invasive method of evaluation into internal joint and muscle force function, as well as athletic performance and neuromuscular coordination. While these are beneficial, current musculoskeletal models lack the ability to give insight into the mechanical stimuli required to prevent physiological deconditioning (i.e., muscle atrophy, bone decalcification, and poor cardiovascular health). This is seen on the International Space Station (ISS) as astronauts still experience deconditioning symptoms even with strict exercise regimes. Therefore, this study overviews the development of a multidimensional surface interpolation, which uses a radial basis function derived from the proper orthogonal decomposition). Inputs to the interpolation scheme are biomechanical responses found by solving an inverse dynamic problem using the ISS’s Advanced Resistive Exercise Device model and simulation within OpenSim. A comparison of 41 lower extremity muscle forces between the interpolation and the static optimization analysis in OpenSim was made. Results highlighted a 1.30% average deviation, with errors concentrated at fluctuating data regions in the minimum force production category. The interpolation scheme shows promise in aiding researchers in the development of an optimized singular exercise movement versus the three exercise machines currently used on the ISS.

Musculoskeletal model-based simulation can be a powerful tool in the evaluation of exoskeletons. An ideal exoskeleton model, perfectly aligned with the human joint axes, can be used to co-simulate the human and exoskeleton dynamics. However, human-exoskeleton joint misalignment is commonly observed during the use of an exoskeleton. Using misaligned motion data in the combined human-exoskeleton model can lead to unrealistic results. In this work, we present a new method to align human-exoskeleton models. This was achieved by introducing dummy segments that ensured kinetic alignment between the human and exoskeleton joints without altering the misaligned kinematics. The method was applied on an active lower-limb exoskeleton that assists the elderly in stair negotiation. In a pilot study, motion data of a single subject testing the exoskeleton in stair ascent were recorded using an optical marker-based system. Measured ground reaction force and exoskeleton assistive force were used as inputs in the human-exoskeleton model. The outputs from the model with the dummy segments were compared to those from a model with kinematic constraints and a reference model where the external forces were applied directly to the human model. The results of the knee compression force, knee flexion moment, and activation of vastus lateralis showed good agreement between the dummy segments and reference models. The use of the dummy segments allows the study of aligned kinetics and misaligned kinematics from the same model. The method will be used in a future study to evaluate the exoskeleton with more subjects.

Magnetic Resonance guided Focused Ultrasound (MRgFUS) is a non-invasive technique based on the thermal ablation of a target using high intensity focused ultrasound. MRgFUS treatment applied to brain is challenging due to the skull presence that attenuates ultrasound, leading to heating effects in bone region. In this study, we simulate trans-cranial nonlinear ultrasound propagation considering the detailed structure of bone tissue. We developed a 2D Finite Element (FE) model that mimics the propagation of focused ultrasound through skin, skull and brain tissue. The skull is represented as a three-layered system with two cortical tables packing a layer of trabecular bone. We assume that the space between the concave transducer and tissue is filled by water. Nonlinear ultrasound propagation is determined through Westervelt equation. To control reflection, absorbing layers have been implemented on the boundaries of the domains. The solution of the pressure equation is subsequently coupled with Pennes bioheat equation to determine the temperature distribution in the tissue region. The acoustic pressure, acoustic intensity and temperature distribution are achieved from FE simulation. Highest values of acoustic pressure occur in the focal area and in the bone tissue region. Ablative temperatures, i.e. superior to 55 °C, are achieved in the target zone and at the cortical-trabecular interface. The thermal response in the focal region is in agreement with available literature and allows to validate the model effectiveness. The FE model offers new insights to predict secondary heating effects of ultrasound propagation in the skull region and to improve treatment planning.

The topic of concussion has become controversial due to the recent discoveries of long-term neurodegenerative diseases in former football players related to concussions they sustained earlier in life. In order to prevent concussion, the mechanism of concussive head impacts in football must be understood. However, studying concussion is difficult because of the ethical issues related to studies involving living subjects. Finite element simulations enable researchers to study the relation between predicted injury in the brain and head impact mechanics as surrogates to live subjects. To improve the accuracy of these simulations, this paper developed a methodology to cyclically analyze and improve the process of finite element modelling of head impacts. Case studies of two college football players were analyzed through the following steps: video analysis of representative concussive hits to elucidate head impact mechanics, finite element simulation of head impact to garner biomechanical metrics in the brain, and correlation of biomechanical metrics with neuroimaging metrics. A relationship was found between these metrics while areas of improvement within the accident reconstruction and finite element simulation process were also found. Through acknowledging areas of further work, researchers can continue to develop the process of head impact simulations to eventually use this tool for diagnostic purposes.

Atherosclerotic cardiovascular disease is a silent and common pathology that affects millions of people around the world. Over time, lipids are deposited in the arterial walls reducing blood flow and as a result can lead to several dangerous and life-threatening cardiovascular issues including myocardial infarction. Due to the high mortality caused by this disease, over the years, great progress has been achieved in understanding the blood fow phenomena through extensive experimental and numerical research. Although in vivo and in vitro experiments have played an important role in the validity of new treatment techniques, due to the great progress that has been made in computational power, numerical methods have been frequently applied in atherosclerosis’ research. Typically, blood flow modeling is performed considering the blood as a homogeneous Newtonian or non-Newtonian fluid. However, blood is a complex fluid that contains a mixture of plasma, red blood cells (RBCs), white blood cells, and platelets. Accordingly, in the present study, the dense-discrete phase model (DDPM), which consists of a hybrid Eulerian-Lagrangian method, is used to simulate the presence of RBCs in blood when flowing through a stenotic coronary arteries. Moreover, the same simulation was performed resorting to the mixture model in order to verify if the differences are considerable. In general, it was found that the generation of recirculation zones is intensified in the DDPM model when compared to the mixture model, but in terms of velocity no significant differences were found.

Cancer cell migration between different body parts is the driving force behind cancer metastasis, which causes mortality of patients. Migration of cancer cells often proceeds by penetration through narrow cavities in possibly stiff tissues. In our previous work [12], a model for the evolution of cell geometry is developed, and in the current study we use this model to investigate whether followers among (cancer) cells benefit from leading (cancer) cells during transmigration through micro-channels and cavities. Using Wilcoxon’s signed-rank text on the data collected from Monte Carlo simulations, we conclude that the transmigration time for the stalk cell is significantly smaller than for the leading cell with a p-value less than 0.0001, for the modelling set-up that we have used in this study.

To preserve the haemostasis of the body, a balance between cell formation and cell death has to be maintained. Thus, when necessary, cells divide into two new ones through a complex and controlled process that is mainly dependent on the availability of oxygen and glucose. Numerical methods have been used to study and solve several problems in science and engineering. Regarding biomechanics and the study of cells, these methods were already combined with distinct formulations. Within these methods, it is possible to highlight the meshless methods, which are a novel and efficient approach that emerged as an alternative to traditional methods. However, the available literature is not yet as extensive as these last ones. The aim of this work is to simulate the process of cell proliferation resorting to a new algorithm. This was based on a phenomenological growth law proposed by the authors, dependent on the presence of oxygen and glucose and solved by a meshless method (Radial Point Interpolation Method). In the end, the algorithm leads to satisfactory and coherent results. It was capable of simulating cell proliferation, under different concentration of oxygen and glucose, as expected. Furthermore, the RPIM proved to be a viable option to solve this problem since the results obtained were adequate.

Primary stability in uncemented total hip arthroplasty is often assessed by insertion force, relative micromotions at the interface or subsidence of the femoral stem. This study aimed to simulate implantation process and evaluate the press-fit effect imposed at this stage on the primary stability outcome. The femur was modeled as an elasto-plastic material with element deletion based on cumulation of plastic strain represented as damage. The implantation phase was simulated using a displacement constraint that led to implant self-positioning by frictional contact. A subsequent force-controlled, loading-unloading cycle corresponding to the peak of a gait cycle was applied. Implantation force and work, implant subsidence and relative micromotions during the loading cycle were quantified. The modeling strategy was performed on two levels of hypothetical initial press-fit to capture the extreme influence of press-fit on the subsequent outcomes. The implantation force was 21.5% higher for the high press-fit approach. The implant remained more stable during the loading cycle for low press-fit model based on subsidence values. Micromotions showed also a significant dependency on press-fit. The median values of micromotions were below 150 µm for both press-fit models, which is in good agreement with the literature. This study demonstrates feasibility of simulating the implantation stage and investigating the effect of press-fit on primary stability of an individual total hip arthroplasty planning.

Based on the patient-specific model of a Type B aortic dissection we created a second model with reduced entry and exit tear size. Two sets of simulations were performed for each model: (i) fluid structure interaction (FSI) and (ii) rigid wall simulations. In both simulation modalities we found that alterations in tear size had substantial impact on true to false lumen flow ratios, true and false lumen pressure differences, and loss of systolic pressure along the dissection. Compared to rigid wall simulations, FSI simulations yielded decreased true lumen flow ratios, increased dampening of flow waveforms along the aorta, smaller negative pressure differences in the distal dissection, and smaller systolic pressure drops across entry and exit tears. These results underline the sensitivity of simulation-based quantitative hemodynamics in aortic dissections to tear size and tissue stiffness.

Time-resolved three-dimensional ultrasound combined with speckle tracking algorithms (4D ultrasound) is a noninvasive medical imaging technique that provides full field displacement and strain data of aortic wall motion. We have developed a Finite Element Model Updating (FEMU) approach to determine the individual nonlinear and anisotropic constitutive behavior of healthy and diseased human aortae in vivo based on 4D ultrasound data, diastolic and systolic blood pressure, only.

In an in vitro study by our group, the 95% confidence interval of the random error of local strain measurement by 4D ultrasound was determined to be ±2.1%. The uncertainty of the mean of the local strains was slightly smaller (± 1.6%). However, the random error of the mean in repeated measurements is the possible systematic error of local strain values in a single measurement.

The uniqueness of the parameter identification and the effect of the measurement uncertainty on the identified constitutive behavior were examined in numerical experiments. The results indicate that the solution of the inverse problem is unique with regard to the identified stretch-Cauchy stress curves (R² ≥ 0.978), but non-unique with regard to the parameters of the constitutive equation. These results were confirmed by parameter identification based on strain field data with overlaid random error (R² ≥ 0.912). In contrast, a systematic error of the mean strain (+1.6%) resulted in the identification of a decisively softer material. The numerical verification shows that it is feasible to identify the constitutive behavior of the geometrically irregular aortic wall based on just two load-cases if heterogeneous strain fields are available.

Introduction: The objective of this systematic review was to analyse 3-D printing in orthopaedic surgery, focussing on pre-operative planning, patient specific implants, instruments, and orthoses.

Method: The PRISMA methodology was followed and literature searches were conducted on Medline, Embase and the Cochrane library. MeSH search terms and Boolean operators included ‘3-D printing’, ‘orthopaedic’, ‘pre-operative plan’, ‘implants’, ‘patient specific instruments’ AND ‘orthosis’.

Results: Searches resulted in 36 studies included in the review. The increasing interest in 3-D printing in orthopaedics is reflected in the rise in publications between 2015 and 2020. The most common application, reported by 75% of the studies was the use of 3-D printed anatomical models to aid in pre-operative planning. The models were also utilized for surgical simulation (31%), intraoperative navigation (8%), and patient/family and surgical education (8%). The use of 3-D printing to manufacture patient specific orthoses, implants and instruments was reported in 14%, 11% and 8% of the studies respectively. The advantage of 3-D printing reported most (56%) was the educational and training opportunities the models provided for junior surgeons. Doctor-patient communication and improved consenting was a reported benefit in 28% of the studies. Objective benefits of 3-D printing included reduced operating time (42%), instrumentation time (11%), fluoroscopy time (31%) and intraoperative blood loss (33%).

Conclusion: The literature shows 3-D printing has improved pre-operative planning, allowed for surgical simulation, training and education. These benefits have led to improved operating metrics. There are currently no studies which demonstrate these reported benefits led to improved patient outcomes.

Digital volume correlation (DVC) performed on micro-computed tomography (CT) imagery provides a measurement technique which can measure full-field deformations of loaded osseous tissues. This experimental approach is of interest in the investigation of the failure mechanisms of glenoid implants in total shoulder arthroplasties, as it allows for direct experimental measurement of strains at the bone-cement-implant interface. It is therefore important to understand the methodological limitations of the bone strain measurements made and the inherent uncertainty present in this approach. Micro-CT scans of two cadaveric scapulae from donors who had been treated with shoulder arthroplasty while living were acquired with differing numbers of CT projections under loaded and unloaded conditions. DVC strain measurements were quantified from the unloaded and loaded volumetric images with five distinct sub-volume sizes. The mean absolute error and measurement sensitivity were quantified in the DVC strains as a function of projection count and sub-volume size, establishing relationships between measurement spatial resolution, image quality, and strain measurement error. Observations revealed that with careful selection of DVC spatial resolution and CT projection count, CT acquisition times can be halved with little impact on DVC strain accuracy. Using 3141 CT projection images as the basis of loaded DVC yielded a maximum mean relative error of 587 με, which increased to 689 με using 1570 CT projection imagery. Thus, with careful consideration of the imaging parameters used, DVC can be a useful preclinical evaluation tool to quantify the internal strain within bone-implant constructs.

During the preoperative planning of total hip replacements (THR) the patient’s biomechanical condition is widely neglected. The result can be a suboptimal implant selection and positioning, which may cause muscular dysfunctions, especially for revision THR. Thus, this study provides evidence in the use of musculoskeletal human models (MHM) during preoperative planning to improve postoperative outcomes. Therefore, a patient-specific MHM is placed in the pose of single leg stance. First, gluteus medius muscle activation is simulated in three adapted situations: healthy, after primary THR and after revision THR. In a second step, a parameter study with adjustable implant parameters (neck-shaft angle, femoral offset and antetorsion angle) is executed to investigate the effects on the gluteus medius activation. The gluteus medius activation of all three situations shows a remarkable increase to the status after revision THR, with a high risk of muscular impairments occurring due to muscle weakening. The parameter study demonstrates a considerable influence of the adjustable parameters not only on gluteus medius activation but also on hip joint contact force. Concluding, MHM not only allow the analysis of a patient’s individual biomechanical condition, but also improve biomechanical outcomes by enabling the simulation of different geometrical THR implant parameters during surgery planning to support the surgeon in identifying the optimal implant parameters and positioning from both the muscle function and a kinematical (bony) perspective.

Silent speech interfaces using non-invasive electromyography (EMG) sensors have been utilized to control internet-of-things devices [1] and provide communications in acoustically challenging environments [2]. However, they have yet to be implemented into Augmented Reality displays, an area they can potentially revolutionize as a human-machine interface by offering low-profile and fluid input. This study overviews the development of a silent speech interface that receives and decodes input from subvocalizations recorded by skin surface EMG sensors, to be used to control a heads-up-display built on a Microsoft HoloLens. Measured muscle activation of the anterior cervical region while a subject subvocalized words from a predetermined library were collected. Trials consisting of subvocalized words were parsed for individual subvocalizations to build a dataset for training of a speech recognition model. The speech recognition model based on a one dimensional convolutional neural network employed to classify subvocalized words was built with the Keras application programming interface in Python, using the TensorFlow library. Preliminary results demonstrate effectiveness in classifying commands, with classification accuracies for ten trained models showing promise. Successful classification was achieved with models showing accuracy in the range of 66.6% to 100%. An average word classification accuracy of 82.5% between all models is observed. While all models were trained and tested on the same datasets, the stochastic nature of the model has significant effects on output, with the dropout layer adding artificial noise to training, and the gradient-descent based optimization algorithm adding random variance to the completed model effectiveness.

Sign Language Recognition aims at providing an efficient and accurate mechanism for the recognition of hand gestures made in sign languages and converting them into text and speech. Sign language is a means of communication using bodily movements, especially using the hands and arms. With sign language recognition methods, dialog communication between the deaf and hearing society can become a reality. In this project, we carry out sign language recognition by building 3D convolutional neural network (3DCNN) models that can perform multi-class prediction on input videos containing hand gestures. On detection of the input, both text and speech are generated and presented as output to the user. In addition to this, we also implement real-time video recognition and continuous sign language recognition for multi-word videos. We present a method for recognition of words in three languages – Tamil Sign Language (TSL), Indian Sign Language (ISL), and American Sign Language (ASL), and outperform state-of-the-art alternatives in with accuracies of 97.5%, 99.75% and 98% respectively.

Motor-unit behaviour is central to understanding muscle contractile function. Commonly, in vivo experimental measurements of motor-unit twitches are taken at points biomechanically linked to the muscle. Here, we investigate whether these measurements accurately represent muscle twitch force behaviour by using a 3D, continuum-mechanical masticatory system model. Selected motor-units within the masseters were individually stimulated, and twitch force measured at the occlusal surface (typical for experiments) and at the muscle origin (representing true twitch force). Occlusal measurements underpredicted twitch force across all motor-units. This was due to force (re)distribution caused by motor-unit location and muscle architecture. Additionally, we investigated the assumption of twitch force summation by comparing twitch forces from co-contracting motor-units with linearly summed counterparts. Shear stress comparisons within the muscle revealed a lower lateral force dispersion for the co-contracting case, partly explaining the underpredicted magnitudes of the linearly summed forces. These findings have implications for twitch force characterisation and motor-unit based muscle modelling.

Biomechanical simulations integrating computations of joint contact pressures within optimizations typically use mesh-based contact models. However, these models are usually not continuous, which could make convergence difficult. Moreover, they can be difficult to tackle when using expression graphs to calculate derivatives (e.g., when using automatic differentiation) during the resolution of optimization problems. This is due to the computational need to use branches when dealing with conditionals. This study presents a mesh-based contact model adapted to deal with gradient-based optimizations and automatic differentiation. A tracking example of a knee prosthesis contact, integrated into a full-body muscle-driven model, is presented. Kinematics and dynamics data were tracked during an overground gait trial. The results show that the accuracy of knee contact force estimations is comparable to other studies (RMSE medial contact force = 50.8 N, RMSE lateral = 81.6 N), and the computational time (3 h and 20 min) is acceptable to be run in a conventional PC. Future work is oriented to formulate predictive simulations of muscle-driven full-body models to predict kinematics and dynamics data of subjects post-surgery simultaneously.

This study presents lower limb joint load comparison from subject specific musculoskeletal model simulation (MSK-MS) and direct measurements from instrumented implants on post-operative (PO) patients. A case study was considered for MSK-MS gait analysis of a 40-year-old healthy male with 70 kg and 1.86 m height. Reflective adhesive markers were applied on skin surface of selected anatomical points at right and left lower limbs. Orthostatic and dynamic acquisition on normal gait (NG), stiff-knee gait (SKG) and slow running (SR) was performed from ground reaction forces with two force plates at 2 kHz and trajectories of skin markers with eight-camera system at 100 Hz. Subject specific MSK-MS was performed using AnyGait and morphed Twente Lower Extremity Model (TLEM), matching the size and joint morphology of the stick-figure model. Over-determinate kinematic analysis was performed, and motion equations solved with hard and soft constraints. Representative MSK-MS gait cycles were selected at NG, SKG and SR lower limb joint vertical force components at the hip, the knee, and the ankle normalized to body weight (JFz/BW). Internal joint direct measurements of four PO patients’, 61–83 years, average weight 808 N and 1.71 m height, with telemetric Hip I (4-channel), Hip II (8-channel) and knee (9-channel) instrumented implants were selected from Orthoload database with comparable gait to NG, SKG and SR. Statistical measurements presented similar mean JFz/BW at right/left hip, knee, ankle MSK-MS and asymmetric peak values with dominant NG, SKG and SR different variances (p < 0.05). Direct JFz/BW measures contrasted NG with similar hip and knee mean and variance from SKG and SR with different mean and variance. Peak JFz/BW direct measurements presented higher hip and knee values on SR and NG than SKG, with higher values at the knee than the hip on NG and SKG, and the opposite on SR. Direct JFz/BW measurements presented at the hip and the knee lower values than their corresponding MSK-MS on NG, SKG and SR.

The skin is a living tissue that behaves in a hyperelastic anisotropic way. To model the skin, we propose an accurate, constitutive law called HGO-Yeoh. This model is implemented in a finite element code FER “Finite Element Research” [4] in order to benefit from its tools, including the bi-potential contact method: a method coupling contact and friction which is more efficient than traditional ones. Identification of parameters of the skin-related material is made with an optimization procedure. A tensile test is simulated using the numerical code FER and ANSYS. The results are compared with the experimental data. A simulation of an indentation test using a bipotential contact law is also performed.

The mechanical properties of breast tissues are important for medical diagnose since the stiffness changes with a pathology. The mechanical behavior of breast tissues has been investigated, however the results in literature are not coherent.s The experimental protocols are not standardized and the need of preconditioning is still debatable, which might be one of the reasons for the variety of results. Thus, this study aims to evaluate the elastic properties of normal breast tissues, considering the effect of preconditioning. Indentation tests were performed and hysteresis was observed. The Young’s modulus ranged from \(3.2 \pm 0.4\) kPa to \(8.8 \pm 1.1\) kPa and from \(43.0\pm 5.2\) kPa to \(117.6 \pm 14.2\) kPa in the first and second linear regions, respectively. Moreover, the preconditioning increased the stiffness on the first linear region. Despite being a preliminary study, important considerations about the experimental protocol were obtained.