Seamless Healthcare Monitoring

This chapter provides a tutorial overview and future perspectives on the historical development and latest advances in electrocardiogram (ECG)-related studies with a focus on the three aspects of signal acquisition, data analysis, and practical application. A development procedure from signal acquisition to knowledge formulation is outlined based on the Signal, Data, Information, Knowledge, Wisdom hierarchical model.

Engineering principles and representative methodologies for ECG acquisition are thoroughly surveyed from Einthoven to Holter. To meet the special requirements for seamless healthcare monitoring, the comprehensive investigation on signal acquisition focuses on measurement methodologies applicable to various daily life scenarios. Acquisition methodologies are categorized into three modalities as wearable, attachable, and invisible. The preferred implementation by the dry and noncontact method is highlighted.

Analysis methods for ECG signal and heart rate data are broadly investigated on a short-term basis, that is, beat-to-beat analysis, sometimes in real-time processing mode, and a long-term basis, meaning various temporal scales such as daily, weekly, monthly, seasonal, and even yearly in batch processing mode. Heart rate variability (HRV) analysis methods in temporal, frequency, and nonlinear domains are extensively reviewed. In addition, heart rate turbulence (HRT) for risk stratification and prediction of acute myocardial infarction is briefly discussed.

The range of practical applications is constantly expanding with the discovery of innovative knowledge and understanding through deep mining of indiscernible information from ECG/HRV. The appropriate combination of such informative features in different analysis domains and the efficient presentation of analytical results play an important role in the visualization of physiological significance. Besides the conventional clinical application of ECG in the diagnosis of cardiovascular-related diseases, broader application in the daily healthcare domain, such as menstrual cycle estimation, health condition tracking, lifestyle change detection, biorhythm evaluation, sleep stage classification, and medicinal effect assessment, is also expected.

Future perspectives are viewed with optimism. Modalities for the acquisition of multiple physiological signals and nonphysiological data over a long-term period are becoming a reality to guarantee analytical outcomes that are more reliable and holistic with the maturation of big data infrastructure, platforms, and analytics, as well as the paradigm shift from clinical disease diagnosis to daily preventive healthcare.

The electroencephalogram (EEG) is a widely used non-invasive method for monitoring the brain. It is based upon placing metal electrodes on the scalp which measure the small electrical potentials that arise outside of the head due to neuronal action within the brain. This chapter overviews the fundamental basis of the EEG, the typical signals that are produced and how they are collected and analysed. Significant attention is given to reviewing the state of the art in EEG collection in both electrode designs and instrumentation hardware. In particular, recent developments in ear-EEG and in conformal tattoo electrodes for very long-term monitoring are highlighted. The chapter concludes by overviewing the applications of EEG technology in medical and non-medical domains, demonstrating the emergence of “consumer neuroscience” applications as EEG devices become more available and more readily useable by non-specialist operators.

In recent years, wireless surface electromyography (SEMG) measurement devices that do not restrict the movement of the wearer have attracted interest in the fields of medicine and sport. Through this review, the reader will develop an understanding of current technologies and potential development.

This section is largely divided into two parts that describe the basic characteristics of EMG and SEMG applications. First, the basic components of EMG will be discussed, as they are useful when focusing on SEMG applications. The physiological information that can be produced with EMG, and SEMG measurement and analysis methods, will be described. Second, specific medical conditions such as lower back pain, stroke, epilepsy, and Parkinson’s disease will be discussed, and the measurement of muscle activity and the information that can be extracted from EMG using smart SEMG measurement devices and systems will be examined. We then focus on the applications of SEMG in sports science, as SEMG have been integrated into wearable platforms such as clothing and textiles; however, a number of problems remain in this regard.

Blood pressure is the most important physiological parameter. A cuff-based sphygmomanometer is commonly used but handling needs great care in terms of cuff size, position of cuff, and so on. A simple handling of wearable blood pressure monitor is desired. Currently, watch-type blood pressure monitor is under development. Whereas cuffless blood pressure monitor has been attempted. Either difference of two pulse wave transit time or R wave of ECG corresponding pulse wave is used to estimate in blood pressure based on biomechanical properties. In this chapter, currently available cuff-based sphygmomanometer is reviewed and then the development of cuffless blood pressure is presented.

Ballistocardiogram (BCG) is the record of mechanical forces exerted by the pumping heart. Movement of the blood volume through the heart chambers and ejection to the arteries causes recoil forces of the body which can be detected with appropriate sensors. It represents the rhythmic activity and the normality of the heart. Several types of sensors can be applied to measure BCG. Among those sensors, accelerometers or film-type sensors are representative, which can measure BCG rather easily without attaching sensors directly to the body surface. These sensors have also merit of easily combining into our daily using devices like chairs, weigh scales, and beds. Usually heart rate and heart rate variability are retrieved from BCG for further application. Measured BCGs are widely used for daily healthcare monitoring including sleep evaluation based on its characteristics of unobtrusiveness.

Based on a wearable pulse rate sensor, photoplethysmographic (PPG) devices are small, inexpensive, safe, and easy to handle and use. Consisting of light-emitting diodes (LEDs) and photodetectors, the PPG device offers a reliable means for monitoring pulse and respiratory rates noninvasively. Recent advances in optical technology have facilitated the use of high-intensity green LEDs for PPG, which has increased the adoption of this measurement technique. In this review, we briefly present the history of PPG, including the development of green PPG, and then discuss recent developments in wearable pulse and respiratory rate sensors. The application of PPG in a clinical trial is also discussed.

Ultrasound (US) is a commonly used cardiac imaging tool in clinical practice. It is a noninvasive, sensitive, and reproducible technique for identifying and quantifying subclinical diseases and for evaluating risk of cardiovascular diseases. Portable handheld US devices have become popular. In this section, the principles of US are reviewed, portable handheld US devices are introduced, and some of their applications of point-of-care technologies are presented.

Wearable inertial sensors have been extensively developed in recent years. Inertial sensors, including accelerometers, gyroscopic sensors, and magnetic sensors, can be embedded in parts of the body, such as the trunk, legs, arms, etc., to monitor motion-related human activities. Inertial sensors are the subject of research as well as of clinical trials. Because sensors must have sufficient accuracy and validity, evaluation of sensor signals is of interest. In this chapter, we examine the technical principles of several types of inertial sensors and provide an assessment of these sensors for patient rehabilitation in clinical practice and sport.

Textile sensors have attempted to measure heart rate, respiratory rate, as well as moving performance. The characteristics of conductivity and elasticity of textile are used to measure angle and performance. In this chapter, the principles of motion analysis using conventional optical method, inertial sensors, and textile sensors are briefly presented. In particular, overview of textile sensors in terms of conductivity and elasticity is explained. Finally, the current topics of sensor application are introduced.

Wearable thermometers are popular devices for measuring body temperature during fever, as well as basal temperature in women. They are easy to handle, inexpensive, and accurate and provide continuous recordings. Most wearable thermometers connect to a smartphone or tablet to display data. Many forms of wearable thermometer are available, such as touch, patch, and invisible (radiometry) types. In this review, we describe and discuss currently available wearable thermometers.

The development of wearable chemical sensors is of interest to obtain comprehensive information for health promotion. However, the development of wearable sensors faces many challenges for long-term use, easy handling, response time, accuracy, validity, and reliability. There are several imitations to produce simple wearable sensors, and above these, optical gas sensors are a promising tool. Monitoring of oxygenation by pulse oximeter and components of expired gas by capnometer is a successful technology. In this section, these two sensors are reviewed including sensor principle and limitation of use.

Seamless monitoring of chemical substances plays a key role in long-term health management, forensics, and security applications. Numerous wearable sensors have been developed to detect different kinds of chemical substances, such as gas, ethanol, urine, glucose, DNA, RNA, and so on. Gas emitted by human can be selectively detected by gas sensor and used as a reference in clinical diagnosis. Glucose sensor can detect the exact concentration of glucose in human blood and can also be used for controllable insulin delivery to reduce the pain to diabetic. The trace elements sensor can detect extremely low concentrations of elements in bio-sample that can be acted as an indicator for certain disease. Moreover, biomarker sensor provides an elusive goal for molecular diagnostics with high accuracy, which is an important tool in early preclinical diagnosis. Portable sensors can be used to detect the amount of ethanol in exhaled gas to confirm whether a driver gets drunk driving. The chemical substances are basically divided into four categories: gas/odor, glucose, trace elements, and biomarker. These sensors could be operated by electrochemical reaction, optical detector, and/or immune antigen-antibody reaction. Electrochemical sensors operate by reacting with the substances of interest and producing an electrical signal proportional to the concentration of analyte (such as hydrogen peroxide in body fluids). The purpose of an optical sensor is to measure a physical quantity of light (such as the amount of light that is scattered by analyte), depending on the type of sensor, and convert the readout to an integrated device to display. Immunosensors can be operated by immunochemical reaction in which antibody immobilized on the solid-state devices can couple with desired analyte (antigens) to produce a transducer signal that can be detected by electrochemical or optical device. Based on the high selectivity of antigen-antibody reaction, the immunosensors can be used for accurate and fast detection of certain biomarker.

This chapter provides an introduction to the field of automatic dietary monitoring (ADM) that intends to derive diet-related behaviour information from unobtrusive sensors and data analysis algorithms. A conceptual gap found in most literature reviews on the relation of physiology and dietary activities is filled. A consistent knowledge-based physiological model for dietary activities is presented. A biomedical approach is adopted to retrieve phenomenological insights of the food preparation, intake, and digestion processes. A taxonomy of dietary activities and a literature review of wearable sensing approaches and dietary dimensions across all dietary activities are also presented.

We begin by briefly introducing the basics of the most frequently used sensors in nowadays wearables targeting a profiling of human physical activity: inertial, biopotential, bioimpedance, and optical sensors. The backbone of the analysis is given to human kinetics and cardiac activity, which are explored in depth in the context of activity profiling in the following sections. Then, an overview of systems for assessing the energy expenditure, calorie consumption, and recovery is presented. Finally, a framework for scientifically evaluating the accuracy of the individual systems is presented.