Physiological Signal Processing Research Articles

Deep learning for processing electromyographic signals: A taxonomy-based survey

Deep Learning (DL) has been recently employed to build smart systems that perform incredibly well in a wide range of tasks, such as image recognition, machine translation, and self-driving cars. In several fields the considerable improvement in the computing hardware and the increasing need for big data analytics has boosted DL work. In recent years physiological signal processing has strongly benefited from deep learning. In general, there is an exponential increase in the number of studies concerning the processing of electromyographic (EMG) signals using DL methods. This phenomenon is mostly explained by the current limitation of myoelectric controlled prostheses as well as the recent release of large EMG recording datasets, e.g. Ninapro. Such a growing trend has inspired us to seek and review recent papers focusing on processing EMG signals using DL methods. Referring to the Scopus database, a systematic literature search of papers published between January 2014 and March 2019 was carried out, and sixty-five papers were chosen for review after a full text analysis. The bibliometric research revealed that the reviewed papers can be grouped in four main categories according to the final application of the EMG signal analysis: Hand Gesture Classification, Speech and Emotion Classification, Sleep Stage Classification and Other Applications. The review process also confirmed the increasing trend in terms of published papers, the number of papers published in 2018 is indeed four times the amount of papers published the year before. As expected, most of the analyzed papers (≈60 %) concern the identification of hand gestures, thus supporting our hypothesis. Finally, it is worth reporting that the convolutional neural network (CNN) is the most used topology among the several involved DL architectures, in fact, the sixty percent approximately of the reviewed articles consider a CNN.

Interoceptive Awareness Is Negatively Related to the Exteroceptive Manipulation of Bodily Self-Location.

The perception of being located within one’s body (i.e., bodily self-location) is an essential feature of everyday self-experience. However, by manipulating exteroceptive input, healthy participants can easily be induced to perceive themselves as being spatially dislocated from their physical bodies. It has previously been suggested that interoception, i.e., the processing of inner physiological signals, contributes to the stability of body representations; however, this relationship has not previously been tested for different dimensions of interoception and bodily self-location. In the present study, using an advanced automatized setup, we systematically manipulated participants’ perspective of their own body (first- vs third-person perspective) as well as the synchrony of visuotactile stimulation (synchronous vs asynchronous). The malleability of bodily self-location was assessed using a questionnaire targeting in-body and out-of-body experiences. Participants also performed a heartbeat discrimination task to assess their interoceptive accuracy (behavioral performance), interoceptive sensibility (confidence in their interoceptive abilities), and interoceptive awareness (meta-cognitive representation of interoceptive signals). Bodily self-location was significantly influenced by perspective, with third-person perspective being associated with stronger out-of-body experiences compared to first-person perspective. Furthermore, there was a significant perspective × stimulation interaction, with subsequent analyses showing that participants reported out-of-body experiences particularly under third-person perspective combined with synchronous visuotactile stimulation. Correlation and regression analyses revealed that meta-cognitive interoceptive awareness was specifically and negatively related to the exteroceptively mediated malleability of body experiences. These results indicate that the perception of the self being located within one’s body relies on the interaction of exteroceptive input and higher-order interoceptive abilities. This has implications for theoretical considerations about the bodily self in health as well as for the understanding of disturbed bodily self-processing in clinical contexts.

Green Tea Polyphenols Promote Functional Recovery from Peripheral Nerve Injury in Rats.

BackgroundPeripheral nerve injury (PNI) is a common and progressive disorder with sensory and motor deficits in the peripheral nervous system (PNS). Treatment is difficult, with unfavorable prognosis. Green tea polyphenols (GTPs) exert neuroprotective effects on regeneration of the central nervous system (CNS). However, the effects of GTPs on functional recovery of the PNS have not been fully characterized. Consequently, the present study investigated the effects of GTPs on nerve regeneration of rats with PNI.Material/MethodsThe model of PNI was established in rats by sciatic nerve injury (SNI). Adult male Wistar rats with SNI were randomly divided into a vehicle group and a GTPs group. The compound muscle action potential (CMAP) of rat sciatic nerves (SN) was measured using the CM6240 physiological signal acquisition and processing system. The wet weight of the triceps muscle was determined using an analytical balance. The number of myelinated nerve fibers was counted under an optical microscope. Ultrastructure of the regenerated nerves in SN was observed by transmission electron microscopy. The mRNA and protein expression of nerve growth factor (NGF), growth-associated protein-43 (GAP-43), neurofilament 200 (NF200), and myelin-associated glycoprotein (MAG) in SN stumps were measured by real-time quantification PCR (RT-qPCR) and Western blot, respectively.ResultsIn rats with SNI, GTPs relieved the adhesion between nerve anastomosis and surrounding tissues, and significantly increased nerve conduction velocity, wet weight of the triceps muscle, and development and axonal regeneration of myelinated nerve fibers. Moreover, GTPs promoted the mRNA and protein expressions of NGF, GAP-43, NF200, and MAG in SN stumps.ConclusionsGTPs promotes nerve regeneration in rats with SNI.

Autonomic Activity and Surgical Flow Disruptions in Healthcare Providers during Cardiac Surgery.

Cardiac surgery represents a complex sociotechnical environment relying on a combination of technical and non-technical team-based expertise. Surgical flow disruptions (SFDs) may be influenced by a variety of sources, including social, environmental, and emotional factors affecting healthcare providers (HCPs). Many of these factors can be readily observed, except for emotional factors (i.e. distress), which represents an underappreciated yet critical source of SFDs. The aim of this study was to demonstrate the sensitivity of autonomic activity metrics to detect an SFD during cardiac surgery. We integrated heart rate variability (HRV) analysis with observation-based annotations to allow data triangulation. Following a critical medication administration error by the anesthesiologist in-training, data sources were consulted to identify events precipitating this near-miss event. Using pyphysio, an open-source physiological signal processing package, we analyzed the attending anesthesiologists' HRV, specifically the low frequency (LF) power, high frequency (HF) power, LF/HF ratio, standard deviation of normal-to-normal (SDNN), and root mean square of the successive differences (RMSSD) as indicators of ANS activity. A heightened SNS response in the attending anesthesiologists' physiological arousal was observed as elevations in LF power and LF/HF ratio, as well as depressions in HF power, SDNN, and RMSSD prior to the near-miss event. The attending anesthesiologist subjectively confirmed a state of high distress induced by task-irrelevant environmental factors during this time. Qualitative analysis of audio/video recordings objectively revealed that the autonomic nervous system (ANS) activation detected was temporally associated with an argument over operating room management. This study confirms that it is possible to recognize detrimental psychophysiological influences in cardiac surgery procedures via advanced HRV analysis. To our knowledge, ours is the first such case demonstrating ANS activity coinciding with strong self-reported emotion during live surgery using HRV. Despite extensive experience in the cardiac OR, transient but intense emotional changes may have the potential to disrupt attention processes in even the most experienced HCP. A primary implication of this work is the possibility to detect real-time ANS activity, which could enable personalized interventions to proactively mitigate downstream adverse events. Additional studies on our large database of surgical cases are underway and new studies are actively being planned to confirm this preliminary observation.

Single-loop sigma delta modulator design and verification for cognitive IoT applications

PurposeA third-order discrete time sigma delta modulator (SDM) is proposed with optimum performance by addressing instability and power dissipations issues, and a novel SDM architecture is designed and verified in behavioural modelling in MATLAB/SIMULINK environment. Simulation results show that performance parameters of proposed modulator achieved SNR of 105.41 dB, SNDR of 101.96 dB and DR of 17 bits for the signal bandwidth of 20 kHz.Design/methodology/approachThis paper describes single-loop SDM design with optimum selection of integrator weights for physiological signal processing in IoT applications.FindingsThe proposed discrete time modulator designed with 1-bit quantizer and optimum oversampling ratio proved as power efficient. Integrator scaling coefficients are generated in LabVIEW environment for pure third-order noise shaping.Originality/valueThis paper contains the novelty in the work, and it is suitable for cognitive Internet of Things applications.

A wearable device for health management detection of multiple physiological parameters based on ZigBee wireless networks

Abstract Detecting physiological signals in the daily life of the population can effectively grasp the health status of the people in the area, and provide an important reference for the diagnosis and prevention of chronic diseases. This paper demonstrates that ZigBee technology meets the requirements of this system. A wireless sensor network system of human physiological parameters based on Zig Bee protocol is built. This wearable device consists of smart hardware and wearables. Intelligent hardware consists of multiple physiological signal detection modules and core processing platforms, and is mainly responsible for the collection and signal processing of multiple physiological signals. Based on the system platform, the entire system was tested for joint debugging. The system's wireless transmission stability and power consumption were tested, and various physiological parameters and motion signals were collected and tested. The test results verify the effectiveness of this system.

A Wavelet Adaptive Cancellation Algorithm Based on Multi-Inertial Sensors for the Reduction of Motion Artifacts in Ambulatory ECGs.

Wearable electrocardiogram (ECG) devices are universally used around the world for patients who have cardiovascular disease (CVD). At present, how to suppress motion artifacts is one of the most challenging issues in the field of physiological signal processing. In this paper, we propose an adaptive cancellation algorithm based on multi-inertial sensors to suppress motion artifacts in ambulatory ECGs. Firstly, this method collects information related to the electrode motion through multi-inertial sensors. Then, the part that is not related to the electrode motion is removed through wavelet transform, which improves the correlation of the reference input signal. In this way, the ability of the adaptive cancellation algorithm to remove motion artifacts is improved in the ambulatory ECG. Subsequent experimentation demonstrated that the wavelet adaptive cancellation algorithm based on multi-inertial sensors can effectively remove motion artifacts in ambulatory ECGs.

Activity Recognition Using Wearable Physiological Measurements: Selection of Features from a Comprehensive Literature Study.

Activity and emotion recognition based on physiological signal processing in health care applications is a relevant research field, with promising future and relevant applications, such as health at work or preventive care. This paper carries out a deep analysis of features proposed to extract information from the electrocardiogram, thoracic electrical bioimpedance, and electrodermal activity signals. The activities analyzed are: neutral, emotional, mental and physical. A total number of 533 features are tested for activity recognition, performing a comprehensive study taking into consideration the prediction accuracy, feature calculation, window length, and type of classifier. Feature selection to know the most relevant features from the complete set is implemented using a genetic algorithm, with a different number of features. This study has allowed us to determine the best number of features to obtain a good error probability avoiding over-fitting, and the best subset of features among those proposed in the literature. The lowest error probability that is obtained is 22.2%, with 40 features, a least squares error classifier, and 40 s window length.

Ranking Power Spectra: A Proof of Concept

To characterize the irregularity of the spectrum of a signal, spectral entropy and its variants are widely adopted measures. However, spectral entropy is invariant under the permutation of the power spectrum estimations on a predefined grid. This erases the inherent order structure in the spectrum. To disentangle the order structure and extract meaningful information from raw digital signal, a novel analysis method is necessary. In this paper, we tried to unfold this order structure by defining descriptors mapping real- and vector-valued power spectrum estimation of a signal into a scalar value. The proposed descriptors showed its potential in diverse problems. Significant differences were observed from brain signals and surface electromyography of different pathological/physiological states. Drastic change accompanied by the alteration of the underlying process of signals enables it as a candidate feature for seizure detection and endpoint detection in speech signal. Since the order structure in the spectrum of physiological signal carries previously ignored information, which cannot be properly extracted by existing techniques, this paper takes one step forward along this direction by proposing computationally efficient descriptors with guaranteed information gain. To the best of our knowledge, this is the first work revealing the effectiveness of the order structure in the spectrum in physiological signal processing.

A New Capacitance Multiplier Structure with High Multiplication Factor for Ultra-Low-Frequency Filter in Biomedical Applications

A novel capacitance multiplier is proposed to implement an ultra-low-frequency filter for physiological signal processing in biomedical applications. With the proposed multiplier, a simple first-order low-pass filter achieves a [Formula: see text]3-dB frequency of 33.4[Formula: see text]μHz with a 1-pF capacitance and a 20[Formula: see text]k[Formula: see text] resistance. This corresponds to a multiplication factor of as large as [Formula: see text]. By changing the controlling terminal, the [Formula: see text]3-dB frequency can be tuned in a wide range of 33.4[Formula: see text]μHz–6.3[Formula: see text]kHz.

Real-Time Driver Drowsiness Detection Using Wavelet Transform and Ensemble Logistic Regression

Drowsy-driver-related accidents has increased in recent years. Research and systems development aim to reduce traffic-accident-related injuries and fatalities. These potentially life-saving systems must operate in a timely manner with the highest precision. In the past two decades, researchers proposed method based on driving pattern changes, driver body position, and physiological signal processing patterns. There is a focus on human physiological signals, specifically the electrical signals from the heart and brain. In this paper, we are presenting an alternative method to determine and quantify driver drowsiness levels using a physiological signal that was collected in a non-intrusive method. This methodology utilizes heart rate variation (HRV), electrocardiogram (ECG), and machine learning for drowsiness detection. Thirty subjects were recruited and ECG data was collected as each subject drifted off to sleep and while sleeping for a duration of between four and eight hours of normal sleep. After using the continuous wavelet transform for the feature extraction, a new feature selection was executed using ensemble logistic regression (ELR), which achieved an average accuracy of 92.5% using data acquired from thirty subjects in an average of 21 s. Successful application of this drowsiness detection method may help prevent traffic accidents.

Machine learning approaches for boredom classification using EEG

Recently, commercial physiological sensors and computing devices have become cheaper and more accessible, while computer systems have become increasingly aware of their contexts, including but not limited to users’ emotions. Consequently, many studies on emotion recognition have been conducted. However, boredom has received relatively little attention as a target emotion due to its diverse nature. Moreover, only a few researchers have tried classifying boredom using electroencephalogram (EEG). In this study, to perform this classification, we first reviewed studies that tried classifying emotions using EEG. Further, we designed and executed an experiment, which used a video stimulus to evoke boredom and non-boredom, and collected EEG data from 28 Korean adult participants. After collecting the data, we extracted its absolute band power, normalized absolute band power, differential entropy, differential asymmetry, and rational asymmetry using EEG, and trained these on three machine learning algorithms: support vector machine, random forest, and k-nearest neighbors (k-NN). We validated the performance of each training model with 10-fold cross validation. As a result, we achieved the highest accuracy of 86.73% using k-NN. The findings of this study can be of interest to researchers working on emotion recognition, physiological signal processing, machine learning, and emotion-aware system development.

Cancellation of motion artifacts in ambulatory ECG signals using TD-LMS adaptive filtering techniques

Wearable electrocardiogram (ECG) measurement systems have been widely used in patients with CVD (Cardiovascular Disease) which can be worn in daily lives. However, currently the main problem is motion artifact interference, and reducing motion artifacts (MA) is one of the most challenging problems encountered in the filtering and processing of physiological signals. In this paper, by analyzing the spectral energy changes during the input process of motion artifacts, a cosine transform LMS adaptive cancellation algorithm (DCT-LMS) implementation is proposed aiming to remove the motion artifacts from the ECG. In order to study the performance of the algorithm and effectively remove the motion artifacts in the ECG signal, this thesis collects ECG signals of people's daily activities from fabric-based chest straps with dry electrodes. It verifies the classic LMS adaptive elimination algorithm and the normalized one. Besides, two LMS adaptive cancellation algorithms based on sine and cosine transform are compared. The simulation and experimental results show that the cosine-based adaptive algorithm is superior to the classical LMS algorithm in eliminating high-amplitude motion artifact noise of ECG.

Research of electroencephalography representational emotion recognition based on deep belief networks

In recent years, with the rapid development of machine learning techniques,the deep learning algorithm has been widely used in one-dimensional physiological signal processing. In this paper we used electroencephalography (EEG) signals based on deep belief network (DBN) model in open source frameworks of deep learning to identify emotional state (positive, negative and neutrals), then the results of DBN were compared with support vector machine (SVM). The EEG signals were collected from the subjects who were under different emotional stimuli, and DBN and SVM were adopted to identify the EEG signals with changes of different characteristics and different frequency bands. We found that the average accuracy of differential entropy (DE) feature by DBN is 89.12%±6.54%, which has a better performance than previous research based on the same data set. At the same time, the classification effects of DBN are better than the results from traditional SVM (the average classification accuracy of 84.2%±9.24%) and its accuracy and stability have a better trend. In three experiments with different time points, single subject can achieve the consistent results of classification by using DBN (the mean standard deviation is1.44%), and the experimental results show that the system has steady performance and good repeatability. According to our research, the characteristic of DE has a better classification result than other characteristics. Furthermore, the Beta band and the Gamma band in the emotional recognition model have higher classification accuracy. To sum up, the performances of classifiers have a promotion by using the deep learning algorithm, which has a reference for establishing a more accurate system of emotional recognition. Meanwhile, we can trace through the results of recognition to find out the brain regions and frequency band that are related to the emotions, which can help us to understand the emotional mechanism better. This study has a high academic value and practical significance, so further investigation still needs to be done.

Biosensor-based feature extraction and physiological parameters measurement for biomedical applications

Physiological signal processing has significantly increased recently among the biomedical researchers for developing wearable standalone devices. Wearable devices are essential in hospital which is used for various biomedical applications such as heart rate and blood pressure measurement. In noisy environment, producing an algorithm for the physiological signal processing and feature extraction is an essential task which depends on the physiological conditions. In this paper, the software-based algorithms are developed for physiological signal processing and physiological parameter measurement. This paper presents biosensor-based real-time acquisition, processing and peak detection of ECG and PPG signals for heart rate, pulse rate and blood pressure measurement. ECG and PPG amplifiers were designed and the real-time signals are acquired through analogue devices and processed using LabVIEW and MatLab environment. The advantages of proposed work are very simple, low cost, easy integration with programming environment and gives continuous monitoring of physiological signals.

Detection of epileptic dysfunctions in EEG signals using Hilbert vibration decomposition

Abstract Epilepsy is a neurological brain dysfunction that is manifested by recrudescent seizures. Due to high temporal resolution, brain activities recorded by electroencephalography (EEG) are commonly used for localization of seizures and identification of epileptic dysfunctions. However, it is often time consuming and challenging to detect EEG seizures using conventional Fourier-based methods and manual interpretation due to nonlinear and nonstationary dynamics of EEG. In this paper, we propose a new framework based on Hilbert vibration decomposition (HVD) for discriminating normal and epileptic EEG recordings. HVD exploits Hilbert transform presentation of instantaneous frequency and extracts monocomponents that have distinctive time-varying amplitudes and instantaneous frequencies from nonstationary signals. The proposed method employs estimated instantaneous frequencies of largest energy components as features given to least squares support vector machine (LS-SVM) for recognizing epileptic seizures and is shown to be appealing for real time physiological signal processing applications due to its reduced computational complexity. Test results on a benchmark EEG data set achieved 97.66% classification accuracy and area of 0.9914 under the receiver operating characteristics (ROC) curve using the delta, theta and alpha rhythms.

A Multimodal Perception Framework for Users Emotional State Assessment in Social Robotics

In this work, we present an unobtrusive and non-invasive perception framework based on the synergy between two main acquisition systems: the Touch-Me Pad, consisting of two electronic patches for physiological signal extraction and processing; and the Scene Analyzer, a visual-auditory perception system specifically designed for the detection of social and emotional cues. It will be explained how the information extracted by this specific kind of framework is particularly suitable for social robotics applications and how the system has been conceived in order to be used in human-robot interaction scenarios.

Helmholtz-Rayleigh Interdisciplinary Silver Medal in Psychological and Physiological Acoustics, Speech Communication, and Signal Processing in Acoustics: Blake S. Wilson

The Silver Medal is presented to individuals, without age limitation, for contributions to the advancement of science, engineering, or human welfare through the application of acoustic principles, or through research accomplishment in acoustics.

A hierarchical classification method for automatic sleep scoring using multiscale entropy features and proportion information of sleep architecture

BackgroundSleep scoring is a critical step in medical researches and clinical applications. Traditional visual scoring method is based on the processing of physiological signals, such as electroencephalogram (EEG), electrooculogram (EOG) and electromyogram (EMG), which is a time consuming and subjective procedure. It is an urgent task to develop an effective method for automatic sleep scoring. MethodThis paper presents a hierarchical classification method for automatic sleep scoring by combining multiscale entropy features with the proportion information of the sleep architecture. Based on a three-layer classification scheme, sleep is categorized into five stages (Awake, S1, S2, SWS and REM). Specifically, the first layer is a standard SVM which performs classification between Awake and Sleep, while the second and third layers are implemented by combining probabilistic output SVM with proportion-based clustering. Multiscale entropy (MSE) from electroencephalogram (EEG) is extracted to represent signal characteristics in multiple temporal scales. ResultsThe proposed method is evaluated with 20 sleep recordings, including 10 subjects with mild difficulty falling asleep and 10 healthy subjects. The overall accuracy of the proposed method is 91.4%. Compared with traditional methods, the classification accuracy of the proposed method is more balanced and the global performance is much better. The dataset includes both healthy subjects and subjects with sleep disorders, which means the presented method has good generalization capacity. Experimental results demonstrate the feasibility of the attempt to introduce proportion information into automatic sleep scoring.

Assessing ECG signal quality indices to discriminate ECGs with artefacts from pathologically different arrhythmic ECGs

False and non-actionable alarms in critical care can be reduced by developing algorithms which assess the trueness of an arrhythmia alarm from a bedside monitor. Computational approaches that automatically identify artefacts in ECG signals are an important branch of physiological signal processing which tries to address this issue. Signal quality indices (SQIs) derived considering differences between artefacts which occur in ECG signals and normal QRS morphology have the potential to discriminate pathologically different arrhythmic ECG segments as artefacts. Using ECG signals from the PhysioNet/Computing in Cardiology Challenge 2015 training set, we studied previously reported ECG SQIs in the scientific literature to differentiate ECG segments with artefacts from arrhythmic ECG segments. We found that the ability of SQIs to discriminate between ECG artefacts and arrhythmic ECG varies based on arrhythmia type since the pathology of each arrhythmic ECG waveform is different. Therefore, to reduce the risk of SQIs classifying arrhythmic events as noise it is important to validate and test SQIs with databases that include arrhythmias. Arrhythmia specific SQIs may also minimize the risk of misclassifying arrhythmic events as noise.