Trends in biomedical signal feature extraction

A convolutional autoencoder framework for ECG signal analysis

Electrocardiographic (ECG) signals are used to evaluate heart activity and to identify disease-related anomalies. Reliable support systems are useful for analyzing ECG signals, for instance, in long-term data acquisition and evaluation (e.g., 24-hour holter recording) or to support physicians in reading ECGs. Analysis of time varying signals may be done by using autoencoders (AEs) deep neural networks. AE specialized for signal data, named Convolutional Autoencoder (CAE), showed the best performances in the analysis of ECG signals. This paper presents a CAE-based framework for ECG signal analysis and anomaly identification. The trained phase is performed on synthetic data signals. The trained neural network obtained is used for the detection of anomalies in ECG signals. The trained framework has been tested on 12 lead ECG signals on a benchmark dataset and applied in scenarios where anomalies are related to cardiological risks and pathologies. The results show interesting results in automatic anomaly detection to support physicians in the decision process. The results show that the CAE-based framework is able to identify anomalies in ECG signals with a ROC AUC of 97.82% on simulated test set and a ROC AUC of 99.75% using a real test set. Finally, the proposed method has been enriched by means of reconstruction error based explainability modules and time-windows based preprocessing modules. Explainability results have been validated using abnormalities annotated by a cardiologist as ground truth and compared with explainations results. System with both code and data, is available at https://github.com/UgoLomoio/ECG_DSS_CAE.

Integrating Remote Photoplethysmography and Machine Learning on Multimodal Dataset for Noninvasive Heart Rate Monitoring

Non-contact heart monitoring is crucial in advancing telemedicine, fitness tracking, and mass screening. Remote photoplethysmography (rPPG) is a non-contact technique to obtain information about heart pulse by analyzing the changes in the light intensity reflected or absorbed by the skin during the blood circulation cycle. However, this technique is sensitive to environmental lightning and different skin pigmentation, resulting in unreliable results. This research presents a multimodal approach to non-contact heart rate estimation by combining facial video and physical attributes, including age, gender, weight, height, and body mass index (BMI). For this purpose, we collected local datasets from 60 individuals containing a 1 min facial video and physical attributes such as age, gender, weight, and height, and we derived the BMI variable from the weight and height. We compare the performance of two machine learning models, support vector regression (SVR) and random forest regression on the multimodal dataset. The experimental results demonstrate that incorporating a multimodal approach enhances model performance, with the random forest model achieving superior results, yielding a mean absolute error (MAE) of 3.057 bpm, a root mean squared error (RMSE) of 10.532 bpm, and a mean absolute percentage error (MAPE) of 4.2% that outperforms the state-of-the-art rPPG methods. These findings highlight the potential for interpretable, non-contact, real-time heart rate measurement systems to contribute effectively to applications in telemedicine and mass screening.

Artificial intelligence in healthcare: a mastery

There is a vast development of artificial intelligence (AI) in recent years. Computational technology, digitized data collection and enormous advancement in this field have allowed AI applications to penetrate the core human area of specialization. In this review article, we describe current progress achieved in the AI field highlighting constraints on smooth development in the field of medical AI sector, with discussion of its implementation in healthcare from a commercial, regulatory and sociological standpoint. Utilizing sizable multidimensional biological datasets that contain individual heterogeneity in genomes, functionality and milieu, precision medicine strives to create and optimize approaches for diagnosis, treatment methods and assessment. With the arise of complexity and expansion of data in the health-care industry, AI can be applied more frequently. The main application categories include indications for diagnosis and therapy, patient involvement and commitment and administrative tasks. There has recently been a sharp rise in interest in medical AI applications due to developments in AI software and technology, particularly in deep learning algorithms and in artificial neural network (ANN). In this overview, we enlisted the major categories of issues that AI systems are ideally equipped to resolve followed by clinical diagnostic tasks. It also includes a discussion of the future potential of AI, particularly for risk prediction in complex diseases, and the difficulties, constraints and biases that must be meticulously addressed for the effective delivery of AI in the health-care sector.

Machine Learning Models for Tracking Blood Loss and Resuscitation in a Hemorrhagic Shock Swine Injury Model

Hemorrhage leading to life-threatening shock is a common and critical problem in both civilian and military medicine. Due to complex physiological compensatory mechanisms, traditional vital signs may fail to detect patients' impending hemorrhagic shock in a timely manner when life-saving interventions are still viable. To address this shortcoming of traditional vital signs in detecting hemorrhagic shock, we have attempted to identify metrics that can predict blood loss. We have previously combined feature extraction and machine learning methodologies applied to arterial waveform analysis to develop advanced metrics that have enabled the early and accurate detection of impending shock in a canine model of hemorrhage, including metrics that estimate blood loss such as the Blood Loss Volume Metric, the Percent Estimated Blood Loss metric, and the Hemorrhage Area metric. Importantly, these metrics were able to identify impending shock well before traditional vital signs, such as blood pressure, were altered enough to identify shock. Here, we apply these advanced metrics developed using data from a canine model to data collected from a swine model of controlled hemorrhage as an interim step towards showing their relevance to human medicine. Based on the performance of these advanced metrics, we conclude that the framework for developing these metrics in the previous canine model remains applicable when applied to a swine model and results in accurate performance in these advanced metrics. The success of these advanced metrics in swine, which share physiological similarities to humans, shows promise in developing advanced blood loss metrics for humans, which would result in increased positive casualty outcomes due to hemorrhage in civilian and military medicine.

Data augmentation for generating synthetic electrogastrogram time series

To address an emerging need for large number of diverse datasets for rigor evaluation of signal processing techniques, we developed and evaluated a new method for generating synthetic electrogastrogram time series. We used electrogastrography (EGG) data from an open database to set model parameters and statistical tests to evaluate synthesized data. Additionally, we illustrated method customization for generating artificial EGG time series alterations caused by the simulator sickness. Proposed data augmentation method generates synthetic EGG data with specified duration, sampling frequency, recording state (postprandial or fasting state), overall noise and breathing artifact injection, and pauses in the gastric rhythm (arrhythmia occurrence) with statistically significant difference between postprandial and fasting states in > 70% cases while not accounting for individual differences. Features obtained from the synthetic EGG signal resembling simulator sickness occurrence displayed expected trends. The code for generation of synthetic EGG time series is not only freely available and can be further customized to assess signal processing algorithms but also may be used to increase data diversity for training artificial intelligence (AI) algorithms. The proposed approach is customized for EGG data synthesis but can be easily utilized for other biosignals with similar nature such as electroencephalogram.

Exploring explainable AI features in the vocal biomarkers of lung disease

This review delves into the burgeoning field of explainable artificial intelligence (XAI) in the detection and analysis of lung diseases through vocal biomarkers. Lung diseases, often elusive in their early stages, pose a significant public health challenge. Recent advancements in AI have ushered in innovative methods for early detection, yet the black-box nature of many AI models limits their clinical applicability. XAI emerges as a pivotal tool, enhancing transparency and interpretability in AI-driven diagnostics. This review synthesizes current research on the application of XAI in analyzing vocal biomarkers for lung diseases, highlighting how these techniques elucidate the connections between specific vocal features and lung pathology. We critically examine the methodologies employed, the types of lung diseases studied, and the performance of various XAI models. The potential for XAI to aid in early detection, monitor disease progression, and personalize treatment strategies in pulmonary medicine is emphasized. Furthermore, this review identifies current challenges, including data heterogeneity and model generalizability, and proposes future directions for research. By offering a comprehensive analysis of explainable AI features in the context of lung disease detection, this review aims to bridge the gap between advanced computational approaches and clinical practice, paving the way for more transparent, reliable, and effective diagnostic tools.

Electromyography as a tool to motion analysis for people with Amyotrophic Lateral Sclerosis: A protocol for a systematic review

Biomechanical analysis of human movement plays an essential role in understanding functional changes in people with Amyotrophic Lateral Sclerosis (ALS), providing information on muscle impairment. Studies suggest that surface electromyography (sEMG) may be able to quantify muscle activity, identify levels of fatigue, assess muscle strength, and monitor variation in limb movement. In this article, a systematic review protocol will analyze the psychometric properties of the sEMG regarding the clinical data on the skeletal muscles of people with ALS. This protocol uses the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodological tool. A specific field structure was defined to reach each phase. Nine scientific databases (PubMed, Web of Science, Embase, Elsevier, IEEE, Google Scholar, SciELO, PEDro, LILACS E CENTRAL) were searched. The framework developed will extract data (i.e. study information, sample information, sEMG information, intervention, and outcomes) from the selected studies using a rigorous approach. The data will be described quantitatively using frequency and trend analysis methods, and heterogeneity between the included studies will be assessed using the I2 test. The risk of bias will be summarized using the most recent prediction model risk of bias assessment tool. Be sure to include relevant statistics here, such as sample sizes, response rates, P values or Confidence Intervals. Be specific (by stating the value) rather than general (eg, "there were differences between the groups"). This protocol will map out the construction of a systematic review that will identify and synthesize the advances in movement analysis of people with ALS through sEMG, using data extracted from articles.

Finding fault types of BLDC motors within UAVs using machine learning techniques

Due to the potential of the Unmanned Aerial Vehicle (UAV), they began to be increasingly used in various fields such as: environment, leisure, health, military, transport, etc. Along with increasing battery storage capacity, the UAVs began to be propulsion by Brushless DC (BLDC) motors. Failure of BLDC motors can lead to loss of control, which can cause accidents. In these conditions, it is necessary to devise methods that can find the defects of the BLDC motors in the UAVs. In this article, the authors propose a novel method to predict BLDC motor defects using machine learning. To maximize the method results, the performance of three machine learning models, K-Nearest Neighbor (KNN), Support Vector Machine (SVM) and Bayesian Network (BN) in predicting the flaws of BLDC motors, have been compared.

Verification of Source Activations in a 3D Brain Model Using 'CLEVER' Algorithm for Mental Arithmetic Conditions

Background: Living conditions are becoming challenging day by day. Mental stress on individuals is increasing due to multiple reasons. As mental stress is a major cause of mental illness, it must be detected at the earliest to prevent serious conditions such as depression and anxiety.

PURPOSE

The focus of this study is to detect the exact location of the source which causes such damage. In this article, we analyse the mental conditions of subjects under a workload of performing mental arithmetic calculations for various frequency bands and plot the topography to understand the areas of active potentials.

METHODS

We propose a Novel Cluster Ensemble Verifier (CLEVER) algorithm, which combines two different techniques: clustering and source localisation. The proposed algorithm is highly efficient in identifying the exact location of the source. It is seen that the topographic plots of the independent component analysis (ICA), which has the maximum percentage of relative variance, correlates to the cluster generated. We are able to give the percentage-wise contribution of every component which is responsible for brain source activation with less time complexity.

RESULTS

Out of 72 subjects, in 67 subjects, 299 out of 433 components originate from the occipital and parietal areas of the brain with a maximum power of 43.5 µv2. As an example, the relative variance of one component is found to be contributing up to 74.03% to source activations. Clusters show similarity across the subjects in the parietal and occipital areas of the brain. The dataset used for experimentation is EEGMAT from Physionet's repository. The computation time for the algorithms is 17.6 ± 3.2 minutes. Conclusion: Findings show that during mental arithmetic calculations, both occipital and parietal areas of the brain are involved. As the data is acquired by orally mentioning the mathematical problem, subjects tend to visualise the numbers while finding the solution, which is reflected in the occipital area of the brain. CLEVER algorithm verifies the origin of the activity in the occipital and parietal areas of the brain.

P300 intention recognition based on phase lag index (PLI)-rich-club brain functional network

Brain-computer interface (BCI) technology based on P300 signals has a broad application prospect in the assessment and diagnosis of clinical diseases and game control. The paper of selecting key electrodes to realize a wearable intention recognition system has become a hotspot for scholars at home and abroad. In this paper, based on the rich-club phenomenon that exists in the process of intention generation, a phase lag index (PLI)-rich-club-based intention recognition method for P300 is proposed. The rich-club structure is a network consisting of electrodes that are highly connected with other electrodes in the process of P300 generation. To construct the rich-club network, this paper uses PLI to construct the brain functional network, calculates rich-club coefficients of the network in the range of k degrees, initially identifies rich-club nodes based on the feature of node degree, and then performs a descending order of betweenness centrality and identifies the nodes with larger betweenness centrality as the specific rich-club nodes, extracts the non-linear features and frequency domain features of Rich-club nodes, and finally uses support vector machine for classification. The experimental results show that the range of rich-club coefficients is smaller with intent compared to that without intent. Validation was performed on the BCI Competition III dataset by reducing the number of channels to 17 and 16 for subject A and subject B, with recognition quasi-departure rates of 96.93% and 94.93%, respectively, and on the BCI Competition II dataset by reducing the number of channels to 17 for subjects, with a recognition accuracy of 95.50%.

Algorithmic detection of sleep-disordered breathing using respiratory signals: a systematic review

Background and Objective. Sleep-disordered breathing (SDB) poses health risks linked to hypertension, cardiovascular disease, and diabetes. However, the time-consuming and costly standard diagnostic method, polysomnography (PSG), limits its wide adoption and leads to underdiagnosis. To tackle this, cost-effective algorithms using single-lead signals (like respiratory, blood oxygen, and electrocardiogram) have emerged. Despite respiratory signals being preferred for SDB assessment, a lack of comprehensive reviews addressing their algorithmic scope and performance persists. This paper systematically reviews 2012-2022 literature, covering signal sources, processing, feature extraction, classification, and application, aiming to bridge this gap and provide future research references.Methods. This systematic review followed the registered PROSPERO protocol (CRD42022385130), initially screening 342 papers, with 32 studies meeting data extraction criteria.Results. Respiratory signal sources include nasal airflow (NAF), oronasal airflow (OAF), and respiratory movement-related signals such as thoracic respiratory effort (TRE) and abdominal respiratory effort (ARE). Classification techniques include threshold rule-based methods (8), machine learning models (13), and deep learning models (11). The NAF-based algorithm achieved the highest average accuracy at 94.11%, surpassing 78.19% for other signals. Hypopnea detection sensitivity with single-source respiratory signals remained modest, peaking at 73.34%. The TRE and ARE signals proved to be reliable in identifying different types of SDB because distinct respiratory disorders exhibited different patterns of chest and abdominal motion.Conclusions. Multiple detection algorithms have been widely applied for SDB detection, and their accuracy is closely related to factors such as signal source, signal processing, feature selection, and model selection.

Advancing spinal cord injury care through non-invasive autonomic dysreflexia detection with AI

This paper presents an AI-powered solution for detecting and monitoring Autonomic Dysreflexia (AD) in individuals with spinal cord injuries. Current AD detection methods are limited, lacking non-invasive monitoring systems. We propose a model that combines skin nerve activity (SKNA) signals with a deep neural network (DNN) architecture to overcome this limitation. The DNN is trained on a meticulously curated dataset obtained through controlled colorectal distension, inducing AD events in rats with spinal cord surgery above the T6 level. The proposed system achieves an impressive average classification accuracy of 93.9% ± 2.5%, ensuring accurate AD identification with high precision (95.2% ± 2.1%). It demonstrates a balanced performance with an average F1 score of 94.4% ± 1.8%, indicating a harmonious balance between precision and recall. Additionally, the system exhibits a low average false-negative rate of 4.8% ± 1.6%, minimizing the misclassification of non-AD cases. The robustness and generalizability of the system are validated on unseen data, maintaining high accuracy, F1 score, and a low false-negative rate. This AI-powered solution represents a significant advancement in non-invasive, real-time AD monitoring, with the potential to improve patient outcomes and enhance AD management in individuals with spinal cord injuries. This research contributes a promising solution to the critical healthcare challenge of AD detection and monitoring.

Analysing the behaviour change of brain regions of methamphetamine abusers using electroencephalogram signals: Hope to design a decision support system

Long-term use of methamphetamine (meth) causes cognitive and neuropsychological impairments. Analysing the impact of this substance on the human brain can aid prevention and treatment efforts. In this study, the electroencephalogram (EEG) signals of meth abusers in the abstinence period and healthy subjects were recorded during eyes-closed and eyes-opened states to distinguish the brain regions that meth can significantly influence. In addition, a decision support system (DSS) was introduced as a complementary method to recognize substance users accompanied by biochemical tests. According to these goals, the recorded EEG signals were pre-processed and decomposed into frequency bands using the discrete wavelet transform (DWT) method. For each frequency band, energy, KS entropy, Higuchi and Katz fractal dimensions of signals were calculated. Then, statistical analysis was applied to select features whose channels contain a p-value less than 0.05. These features between two groups were compared, and the location of channels containing more features was specified as discriminative brain areas. Due to evaluating the performance of features and distinguishing the two groups in each frequency band, features were fed into a k-nearest neighbour (KNN), support vector machine (SVM), multilayer perceptron neural networks (MLP) and linear discriminant analysis (LDA) classifiers. The results indicated that prolonged consumption of meth has a considerable impact on the brain areas responsible for working memory, motor function, attention, visual interpretation, and speech processing. Furthermore, the best classification accuracy, almost 95.8%, was attained in the gamma band during the eyes-closed state.

Acting Emotions: a comprehensive dataset of elicited emotions

Emotions encompass physiological systems that can be assessed through biosignals like electromyography and electrocardiography. Prior investigations in emotion recognition have primarily focused on general population samples, overlooking the specific context of theatre actors who possess exceptional abilities in conveying emotions to an audience, namely acting emotions. We conducted a study involving 11 professional actors to collect physiological data for acting emotions to investigate the correlation between biosignals and emotion expression. Our contribution is the DECEiVeR (DatasEt aCting Emotions Valence aRousal) dataset, a comprehensive collection of various physiological recordings meticulously curated to facilitate the recognition of a set of five emotions. Moreover, we conduct a preliminary analysis on modeling the recognition of acting emotions from raw, low- and mid-level temporal and spectral data and the reliability of physiological data across time. Our dataset aims to leverage a deeper understanding of the intricate interplay between biosignals and emotional expression. It provides valuable insights into acting emotion recognition and affective computing by exposing the degree to which biosignals capture emotions elicited from inner stimuli.

Wearable Sensors as a Preoperative Assessment Tool: A Review

Surgery is a common first-line treatment for many types of disease, including cancer. Mortality rates after general elective surgery have seen significant decreases whilst postoperative complications remain a frequent occurrence. Preoperative assessment tools are used to support patient risk stratification but do not always provide a precise and accessible assessment. Wearable sensors (WS) provide an accessible alternative that offers continuous monitoring in a non-clinical setting. They have shown consistent uptake across the perioperative period but there has been no review of WS as a preoperative assessment tool. This paper reviews the developments in WS research that have application to the preoperative period. Accelerometers were consistently employed as sensors in research and were frequently combined with photoplethysmography or electrocardiography sensors. Pre-processing methods were discussed and missing data was a common theme; this was dealt with in several ways, commonly by employing an extraction threshold or using imputation techniques. Research rarely processed raw data; commercial devices that employ internal proprietary algorithms with pre-calculated heart rate and step count were most commonly employed limiting further feature extraction. A range of machine learning models were used to predict outcomes including support vector machines, random forests and regression models. No individual model clearly outperformed others. Deep learning proved successful for predicting exercise testing outcomes but only within large sample-size studies. This review outlines the challenges of WS and provides recommendations for future research to develop WS as a viable preoperative assessment tool.

Machine Learning for Automated Classification of Abnormal Lung Sounds Obtained from Public Databases: A Systematic Review

Pulmonary auscultation is essential for detecting abnormal lung sounds during physical assessments, but its reliability depends on the operator. Machine learning (ML) models offer an alternative by automatically classifying lung sounds. ML models require substantial data, and public databases aim to address this limitation. This systematic review compares characteristics, diagnostic accuracy, concerns, and data sources of existing models in the literature. Papers published from five major databases between 1990 and 2022 were assessed. Quality assessment was accomplished with a modified QUADAS-2 tool. The review encompassed 62 studies utilizing ML models and public-access databases for lung sound classification. Artificial neural networks (ANN) and support vector machines (SVM) were frequently employed in the ML classifiers. The accuracy ranged from 49.43% to 100% for discriminating abnormal sound types and 69.40% to 99.62% for disease class classification. Seventeen public databases were identified, with the ICBHI 2017 database being the most used (66%). The majority of studies exhibited a high risk of bias and concerns related to patient selection and reference standards. Summarizing, ML models can effectively classify abnormal lung sounds using publicly available data sources. Nevertheless, inconsistent reporting and methodologies pose limitations to advancing the field, and therefore, public databases should adhere to standardized recording and labeling procedures.

Shannon Entropy Used for Feature Extractions of Optical Patterns in the Context of Structural Health Monitoring

A novelty signal processing method is proposed for a technical vision system (TVS). During data acquisition of an optoelectrical signal, part of this is random electrical fluctuation of voltages. Information theory (IT) is a well-known field that deals with random processes. A method based on using of the Shannon Entropy for feature extractions of optical patterns is presented. IT is implemented in structural health monitoring (SHM) to augment the accuracy of optoelectronic signal classifiers for a metrology subsystem of the TVS. To enhance the TVS spatial coordinate measurement performance at real operation conditions with electrical and optical noisy environments to estimate structural displacement better and evaluate its health for a better estimation of structural displacement and the evaluation of its health. Five different machine learning (ML) techniques are used in this work to classify optical patterns captured with the TVS. Linear predictive coding (LPC) and Autocorrelation function (ACC) are for extraction of optical patterns. The Shannon entropy segmentation (SH) method extracts relevant information from optical patterns, and the model's performance can be improved. The results reveal that segmentation with Shannon's entropy can achieve over 95.33%. Without Shannon's entropy, the worst accuracy was 33.33%.

An Explainable Machine-Learning Model for Compensatory Reserve Measurement: Methods for Feature Selection and the Effects of Subject Variability

Tracking vital signs accurately is critical for triaging a patient and ensuring timely therapeutic intervention. The patient's status is often clouded by compensatory mechanisms that can mask injury severity. The compensatory reserve measurement (CRM) is a triaging tool derived from an arterial waveform that has been shown to allow for earlier detection of hemorrhagic shock. However, the deep-learning artificial neural networks developed for its estimation do not explain how specific arterial waveform elements lead to predicting CRM due to the large number of parameters needed to tune these models. Alternatively, we investigate how classical machine-learning models driven by specific features extracted from the arterial waveform can be used to estimate CRM. More than 50 features were extracted from human arterial blood pressure data sets collected during simulated hypovolemic shock resulting from exposure to progressive levels of lower body negative pressure. A bagged decision tree design using the ten most significant features was selected as optimal for CRM estimation. This resulted in an average root mean squared error in all test data of 0.171, similar to the error for a deep-learning CRM algorithm at 0.159. By separating the dataset into sub-groups based on the severity of simulated hypovolemic shock withstood, large subject variability was observed, and the key features identified for these sub-groups differed. This methodology could allow for the identification of unique features and machine-learning models to differentiate individuals with good compensatory mechanisms against hypovolemia from those that might be poor compensators, leading to improved triage of trauma patients and ultimately enhancing military and emergency medicine.

Pseudomonas aeruginosa detection based on droplets incubation using an integrated microfluidic chip, laser spectroscopy, and machine learning

Pseudomonas aeruginosa is an opportunist pathogen responsible for causing several infections in the human body, especially in patients with weak immune systems. The proposed approach reports a novel pathogens detection system based on cultivating microdroplets and acquiring the scattered light signals from the incubated droplets using a microfluidic device. Initially, the microdroplets were generated and incubated to cultivate bacteria inside the microdroplets. The second part of the microfluidic chip is the detection module, embedded with three optical fibers to connect laser light and photosensors. The incubated droplets were reinjected in the detection module and passed through the laser light. The surrounding photosensors were arranged symmetrically at 45° to the flowing channel for acquiring the scattered light signal. The noise was removed from the acquired data, and time-domain waveform features were evaluated. The acquired features were trained using machine learning classifiers to classify P. aeruginosa. The k-nearest neighbors (KNN) showed superior classification performance with 95.6 % accuracy among other classifiers, including logistic regression (LR), support vector machines (SVM), and naïve Bayes (NB). The proposed research was performed to validate the method for pathogens detection with a concentration of 10⁵ CFU/mL. The total duration of 6 h is required to test the sample, including five hours for droplets incubation and one hour for sample preparation and detection using light scattering module. The results indicate that acquiring the light scattering patterns from incubated droplets can detect P. aeruginosa using machine learning classification. The proposed system is anticipated to be helpful as a rapid device for diagnosing pathogenic infections.

Computer-aided diagnosis of autism spectrum disorder from EEG signals using deep learning with FAWT and multiscale permutation entropy features

Autism spectrum disorder (ASD), a neurodevelopment disorder, is characterized by significant difficulties in social interaction and emerges as a major threat to children. Its computer-aided diagnosis used by neurologists improves the detection process and has a favorable impact on patients' health. Currently, a biomarker termed electroencephalography (EEG) is considered as vital tool to detect abnormal electrical activity in the brain. In this context, the present paper brings forth a novel approach for automated diagnosis of ASD from multichannel EEG signals using flexible analytic wavelet transform (FAWT). Firstly, this approach processes the acquired EEG signals with filtering and segmentation into short-duration EEG segments in the range of 5-20 s. These segmented EEG signals are decomposed into five levels using FAWT technique to obtain various sub-bands. Further, multiscale permutation entropy values are extracted from decomposed sub-bands which are used as feature vectors in the present work. Afterwards, these feature vectors are evaluated by traditional machine learning algorithms viz., k-nearest neighbor, logistic regression, support vector machine, and random forest, as well as convolutional neural network (CNN) as deep learning algorithm with different segment durations. The analysis of results reveals that CNN provides maximum accuracy, sensitivity, specificity, and area under the curve of 99.19%, 99.34%, 99.21%, and 0.9997, respectively, for 10 s duration EEG segment to identify ASD patients among healthy individuals. Thus, the proposed CNN architecture would be extremely helpful during diagnostic process of autism disease for neurologists.

Trends in EEG signal feature extraction applications

This paper will focus on electroencephalogram (EEG) signal analysis with an emphasis on common feature extraction techniques mentioned in the research literature, as well as a variety of applications that this can be applied to. In this review, we cover single and multi-dimensional EEG signal processing and feature extraction techniques in the time domain, frequency domain, decomposition domain, time-frequency domain, and spatial domain. We also provide pseudocode for the methods discussed so that they can be replicated by practitioners and researchers in their specific areas of biomedical work. Furthermore, we discuss artificial intelligence applications such as assistive technology, neurological disease classification, brain-computer interface systems, as well as their machine learning integration counterparts, to complete the overall pipeline design for EEG signal analysis. Finally, we discuss future work that can be innovated in the feature extraction domain for EEG signal analysis.

Prediction of exam scores using a multi-sensor approach for wearable exam stress dataset with uniform preprocessing

BACKGROUND

Physiological signals, such as skin conductance, heart rate, and temperature, provide valuable insight into the physiological responses of students to stress during examination sessions.

OBJECTIVE

The primary objective of this research is to explore the effectiveness of physiological signals in predicting grades and to assess the impact of different models and feature selection techniques on predictive performance.

METHODS

We extracted a comprehensive feature vector comprising 301 distinct features from seven signals and implemented a uniform preprocessing technique for all signals. In addition, we analyzed different algorithmic selection features to design relevant features for robust and accurate predictions.

RESULTS

The study reveals promising results, with the highest scores achieved using 100 and 150 features. The corresponding values for accuracy, AUROC, and F1-Score are 0.9, 0.89, and 0.87, respectively, indicating the potential of physiological signals for accurate grade prediction.

CONCLUSION

The findings of this study suggest practical applications in the field of education, where the use of physiological signals can help students cope with exam stress and improve their academic performance. The importance of feature selection and the use of appropriate models highlight the importance of engineering relevant features for precise and reliable predictions.

Limb accelerations during sleep are related to measures of strength, sensation, and spasticity among individuals with spinal cord injury

BACKGROUND

To evaluate the relationship between measures of neuromuscular impairment and limb accelerations (LA) collected during sleep among individuals with chronic spinal cord injury (SCI) to provide evidence of construct and concurrent validity for LA as a clinically meaningful measure.

METHODS

The strength (lower extremity motor score), sensation (summed lower limb light touch scores), and spasticity (categorized lower limb Modified Ashworth Scale) were measured from 40 adults with chronic (≥ 1 year) SCI. Demographics, pain, sleep quality, and other covariate or confounding factors were measured using self-report questionnaires. Each participant then wore ActiGraph GT9X Link accelerometers on their ankles and wrist continuously for 1-5 days to measure LA from movements during sleep. Regression models with built-in feature selection were used to determine the most relevant LA features and the association to each measure of impairment.

RESULTS

LA features were related to measures of impairment with models explaining 69% and 73% of the variance (R²) in strength and sensation, respectively, and correctly classifying 81.6% (F1-score = 0.814) of the participants into spasticity categories. The most commonly selected LA features included measures of power and frequency (frequency domain), movement direction (correlation between axes), consistency between movements (relation to recent movements), and wavelet energy (signal characteristics). Rolling speed (change in angle of inclination) and movement smoothness (median crossings) were uniquely associated with strength. When LA features were included, an increase of 72% and 222% of the variance was explained for strength and sensation scores, respectively, and there was a 34% increase in spasticity classification accuracy compared to models containing only covariate features such as demographics, sleep quality, and pain.

CONCLUSION

LA features have shown evidence of having construct and concurrent validity, thus demonstrating that LA are a clinically-relevant measure related to lower limb strength, sensation, and spasticity after SCI. LA may be useful as a more detailed measure of impairment for applications such as clinical prediction models for ambulation.

Automatic diagnosis of sleep apnea from biomedical signals using artificial intelligence techniques: Methods, challenges, and future works

Apnea is a sleep disorder that stops or reduces airflow for a short time during sleep. Sleep apnea may last for a few seconds and happen for many while sleeping. This reduction in breathing is associated with loud snoring, which may awaken the person with a feeling of suffocation. So far, a variety of methods have been introduced by researchers to diagnose sleep apnea, among which the polysomnography (PSG) method is known to be the best. Analysis of PSG signals is very complicated. Many studies have been conducted on the automatic diagnosis of sleep apnea from biological signals using artificial intelligence (AI), including machine learning (ML) and deep learning (DL) methods. This research reviews and investigates the studies on the diagnosis of sleep apnea using AI methods. First, computer aided diagnosis system (CADS) for sleep apnea using ML and DL techniques along with its parts including dataset, preprocessing, and ML and DL methods are introduced. This research also summarizes the important specifications of the studies on the diagnosis of sleep apnea using ML and DL methods in a table. In the following, a comprehensive discussion is made on the studies carried out in this field. The challenges in the diagnosis of sleep apnea using AI methods are of paramount importance for researchers. Accordingly, these obstacles are elaborately addressed. In another section, the most important future works for studies on sleep apnea detection from PSG signals and AI techniques are presented. Ultimately, the essential findings of this study are provided in the conclusion section.

This article is categorized under:

Computational Methods for Physiological Signal Processing and Data Analysis

Biomedical signal processing and data analysis play pivotal roles in the advanced medical expert system solutions. Signal processing tools are able to diminish the potential artifact effects and improve the anticipative signal quality. Data analysis techniques can assist in reducing redundant data dimensions and extracting dominant features associated with pathological status. Recent computational methods have greatly improved the effectiveness of signal processing and data analysis, to support the efficient point-of-care diagnosis and accurate medical decision-making. This editorial article highlights the research works published in the special issue of Computational Methods for Physiological Signal Processing and Data Analysis. The context introduces three deep learning applications in epileptic seizure detection, human exercise intensity analysis, and lung nodule CT image segmentation, respectively. The article also summarizes the research works on detection of event-related potential in the single-trial electroencephalogram (EEG) signals during the auditory tests, along with the methodology on estimating the generalized exponential distribution parameters using the simulated and real data produced under the Type I generalized progressive hybrid censoring schemes. The article concludes with perspectives and discussions on future trends in biomedical signal processing and data analysis technologies.

Machine Learning Strategies for Low-Cost Insole-Based Prediction of Center of Gravity during Gait in Healthy Males

Whole-body center of gravity (CG) movements in relation to the center of pressure (COP) offer insights into the balance control strategies of the human body. Existing CG measurement methods using expensive measurement equipment fixed in a laboratory environment are not intended for continuous monitoring. The development of wireless sensing technology makes it possible to expand the measurement in daily life. The insole system is a wearable device that can evaluate human balance ability by measuring pressure distribution on the ground. In this study, a novel protocol (data preparation and model training) for estimating the 3-axis CG trajectory from vertical plantar pressures was proposed and its performance was evaluated. Input and target data were obtained through gait experiments conducted on 15 adult and 15 elderly males using a self-made insole prototype and optical motion capture system. One gait cycle was divided into four semantic phases. Features specified for each phase were extracted and the CG trajectory was predicted using a bi-directional long short-term memory (Bi-LSTM) network. The performance of the proposed CG prediction model was evaluated by a comparative study with four prediction models having no gait phase segmentation. The CG trajectory calculated with the optoelectronic system was used as a golden standard. The relative root mean square error of the proposed model on the 3-axis of anterior/posterior, medial/lateral, and proximal/distal showed the best prediction performance, with 2.12%, 12.97%, and 12.47%. Biomechanical analysis of two healthy male groups was conducted. A statistically significant difference between CG trajectories of the two groups was shown in the proposed model. Large CG sway of the medial/lateral axis trajectory and CG fall of the proximal/distal axis trajectory is shown in the old group. The protocol proposed in this study is a basic step to have gait analysis in daily life. It is expected to be utilized as a key element for clinical applications.

Fault Diagnosis of Permanent Magnet DC Motors Based on Multi-Segment Feature Extraction

For permanent magnet DC motors (PMDCMs), the amplitude of the current signals gradually decreases after the motor starts. Only using the signal features of current in a single segment is not conducive to fault diagnosis for PMDCMs. In this work, multi-segment feature extraction is presented for improving the effect of fault diagnosis of PMDCMs. Additionally, a support vector machine (SVM), a classification and regression tree (CART), and the k-nearest neighbor algorithm (k-NN) are utilized for the construction of fault diagnosis models. The time domain features extracted from several successive segments of current signals make up a feature vector, which is adopted for fault diagnosis of PMDCMs. Experimental results show that multi-segment features have a better diagnostic effect than single-segment features; the average accuracy of fault diagnosis improves by 19.88%. This paper lays the foundation of fault diagnosis for PMDCMs through multi-segment feature extraction and provides a novel method for feature extraction.

Microwave breast tumor localization using wavelet feature extraction and genetic algorithm-neural network

PURPOSE

Ultra-Wide Band (UWB) microwave breast cancer detection is a promising new technology for routine physical examination and home monitoring. The existing microwave imaging algorithms for breast tumor detection are complex and the effect is still not ideal, due to the heterogeneity of breast tissue, skin reflection, and fibroglandular tissue reflection in backscatter signals. This study aims to develop a machine learning method to accurately locate breast tumor.

METHODS

A microwave-based breast tumor localization method is proposed by time-frequency feature extraction and neural network technology. First, the received microwave array signals are converted into representative and compact features by 4-level Discrete Wavelet Transform (DWT) and Principal Component Analysis (PCA). Then, the Genetic Algorithm-Neural Network (GA-NN) is developed to tune hyper-parameters of the neural network adaptively. The neural network embedded in the GA-NN algorithm is a four-layer architecture and 10-fold cross-validation is performed. Through the trained neural network, the tumor localization performance is evaluated on four datasets that are created by FDTD simulation method from 2-D MRI-derived breast models with varying tissue density, shape, and size. Each dataset consists of 1000 backscatter signals with different tumor positions, in which the ratio of training set to test set is 9:1. In order to verify the generalizability and scalability of the proposed method, the tumor localization performance is also tested on a 3-D breast model.

RESULTS

For these 2-D breast models with unknown tumor locations, the evaluation results show that the proposed method has small location errors, which are 0.6076 mm, 3.0813 mm, 2.0798 mm, and 3.2988 mm, respectively, and high accuracy, which is 99%, 80%, 94%, and 85%, respectively. Furthermore, the location error and the prediction accuracy of the 3-D breast model are 3.3896 mm and 81%.

CONCLUSIONS

These evaluation results demonstrate that the proposed machine learning method is effective and accurate for microwave breast tumor localization. The traditional microwave-based breast cancer detection method is to reconstruct the entire breast image to highlight the tumor. Compared with the traditional method, our proposed method can directly get the breast tumor location by applying neural network to the received microwave array signals, and circumvent any complicated image reconstruction processing.

A novel method based on matching pursuit decomposition of gait signals for Parkinson's disease, Amyotrophic lateral sclerosis and Huntington's disease detection

BACKGROUND AND OBJECTIVE

An accurate detection of neurodegenerative diseases (NDDs) definitely improves the life of patients and has attracted growing attention.

METHODS

In this paper, a general automatic method for detection of Parkinson's disease (PD), Amyotrophic lateral sclerosis (ALS) and Huntington's disease (HD) has been proposed based on the localized time-frequency information of gait signals. The new main part of the detection method is to obtain a small set of sparse coefficients for the local representation of gait signals with appropriate time and frequency resolution. For this purpose, a hybrid feature set based on sparse matching pursuit decomposition and two sets of nonlinear and linear features has been developed. Then, principal components of the proposed feature have been analyzed using a sparse coding classifier. Results The proposed approach has achieved high average accuracy rates of 93%, 94%, and 97% for PD, ALS, and HD detection, respectively.

CONCLUSIONS

The obtained results have indicated that combination of time and frequency information of the gait signals through adaptive localized window length in MP makes it more efficient in comparison with the existing time, frequency or other time-frequency gait parameters. The great potential of nonlinear sparse features for PD and HD detection and linear ones for ALS detection has also been shown.

Wavelet Power Spectral Domain Functional Principal Component Analysis for Feature Extraction of Epileptic EEGs

Feature extraction plays an important role in machine learning for signal processing, particularly for low-dimensional data visualization and predictive analytics. Data from real-world complex systems are often high-dimensional, multi-scale, and non-stationary. Extracting key features of this type of data is challenging. This work proposes a novel approach to analyze Epileptic EEG signals using both wavelet power spectra and functional principal component analysis. We focus on how the feature extraction method can help improve the separation of signals in a low-dimensional feature subspace. By transforming EEG signals into wavelet power spectra, the functionality of signals is significantly enhanced. Furthermore, the power spectra transformation makes functional principal component analysis suitable for extracting key signal features. Therefore, we refer to this approach as a double feature extraction method since both wavelet transform and functional PCA are feature extractors. To demonstrate the applicability of the proposed method, we have tested it using a set of publicly available epileptic EEGs and patient-specific, multi-channel EEG signals, for both ictal signals and pre-ictal signals. The obtained results demonstrate that combining wavelet power spectra and functional principal component analysis is promising for feature extraction of epileptic EEGs. Therefore, they can be useful in computer-based medical systems for epilepsy diagnosis and epileptic seizure detection problems.

Interpretable Classification of Bacterial Raman Spectra with Knockoff Wavelets

Deep neural networks and other machine learning models are widely applied to biomedical signal data because they can detect complex patterns and compute accurate predictions. However, the difficulty of interpreting such models is a limitation, especially for applications involving high-stakes decision, including the identification of bacterial infections. This paper considers fast Raman spectroscopy data and demonstrates that a logistic regression model with carefully selected features achieves accuracy comparable to that of neural networks, while being much simpler and more transparent. Our analysis leverages wavelet features with intuitive chemical interpretations, and performs controlled variable selection with knockoffs to ensure the predictors are relevant and non-redundant. Although we focus on a particular data set, the proposed approach is broadly applicable to other types of signal data for which interpretability may be important.

Joint angle estimation with wavelet neural networks

This paper presents a wavelet neural network (WNN) based method to reduce reliance on wearable kinematic sensors in gait analysis. Wearable kinematic sensors hinder real-time outdoor gait monitoring applications due to drawbacks caused by multiple sensor placements and sensor offset errors. The proposed WNN method uses vertical Ground Reaction Forces (vGRFs) measured from foot kinetic sensors as inputs to estimate ankle, knee, and hip joint angles. Salient vGRF inputs are extracted from primary gait event intervals. These selected gait inputs facilitate future integration with smart insoles for real-time outdoor gait studies. The proposed concept potentially reduces the number of body-mounted kinematics sensors used in gait analysis applications, hence leading to a simplified sensor placement and control circuitry without deteriorating the overall performance.

A FUSION FRAMEWORK BASED ON CEPSTRAL DOMAIN FEATURES FROM PHONOCARDIOGRAM TO PREDICT HEART HEALTH STATUS

A great increase in the number of cardiovascular cases has been a cause of serious concern for the medical experts all over the world today. In order to achieve valuable risk stratification for patients, early prediction of heart health can benefit specialists to make effective decisions. Heart sound signals help to know about the condition of heart of a patient. Motivated by the success of cepstral features in speech signal classification, authors have used here three different cepstral features, viz. Mel-frequency cepstral coefficients (MFCCs), gammatone frequency cepstral coefficients (GFCCs), and Mel-spectrogram for classifying phonocardiogram into normal and abnormal. Existing research has explored only MFCCs and Mel-feature set extensively for classifying the phonocardiogram. However, in this work, the authors have used a fusion of GFCCs with MFCCs and Mel-spectrogram, and achieved a better accuracy score of 0.96 with sensitivity and specificity scores as 0.91 and 0.98, respectively. The proposed model has been validated on the publicly available benchmark dataset PhysioNet 2016.

Development of Miniaturized Wearable Wristband Type Surface EMG Measurement System for Biometric Authentication

Personal authentication systems employing biometrics are attracting increasing attention owing to their relatively high security compared to existing authentication systems. In this study, a wearable electromyogram (EMG) system that can be worn on the forearm was developed to detect EMG signals and, subsequently, apply them for personal authentication. In previous studies, wet electrodes were attached to the skin for measuring biosignals. Wet electrodes contain adhesives and conductive gels, leading to problems such as skin rash and signal-quality deterioration in long-term measurements. The miniaturized wearable EMG system developed in this study comprised flexible dry electrodes attached to the watch strap, enabling EMG measurements without additional electrodes. In addition, for accurately classifying and applying the measured signal to the personal authentication system, an optimal algorithm for classifying the EMG signals based on a multi-class support vector machine (SVM) model was implemented. The model using cubic SVM achieved the highest personal authentication rate of 87.1%. We confirmed the possibility of implementing a wearable authentication system by measuring the EMG signal and artificial intelligence analysis algorithm presented in this study.

Can the application of certain music information retrieval methods contribute to the machine learning classification of electrocardiographic signals?

The electrocardiogram is traditionally used to diagnose a large number of heart pathologies. Research to improve the readability and classification of cardiac signals includes studies geared toward sonification of the electrocardiographic signal and others involving features related to music processing, such as Mel-frequency cepstral coefficients. In terms of music processing features, this study seeks to use music information retrieval (MIR) features as electrocardiographic signal descriptors. The study compares the discriminatory capability of the introduced features in relation to standard groups such as heart rate variability, wavelet transform, descriptive statistics, Mel coefficients and fractal analysis, evaluated using classification algorithms; the signals analyzed were extracted from public databases. The group of features extracted from wavelet transform and the MIR group showed a high level of discrimination; the best representation of the ECG signals in the study was achieved in most cases by the MIR features. Moreover, a correlation coefficient higher than 0.8 was found between a number of MIR and other feature groups, indicating a likely relationship between the electrocardiographic signals and MIR features. These results suggest the feasibility of representing the analyzed signals by music information retrieval descriptors, giving the potential to consider these electrocardiographic signals as analogues to musical signals.

A study on CNN image classification of EEG Signals represented in 2D and 3D

Objective The novelty of this study consists of the exploration of multiple new approaches of data pre-processing of brainwave signals, wherein statistical features are extracted and then formatted as visual images based on the order in which dimensionality reduction algorithms select them. This data is then treated as visual input for 2D and 3D CNNs which then further extract 'features of features'. Approach Statistical features derived from three electroencephalography datasets are presented in visual space and processed in 2D and 3D space as pixels and voxels respectively. Three datasets are benchmarked, mental attention states and emotional valences from the four TP9, AF7, AF8 and TP10 10-20 electrodes and an eye state data from 64 electrodes. 729 features are selected through three methods of selection in order to form 27x27 images and 9x9x9 cubes from the same datasets. CNNs engineered for the 2D and 3D preprocessing representations learn to convolve useful graphical features from the data. Main results: A 70/30 split method shows that the strongest methods for classification accuracy of feature selection are One Rule for attention state and Relative Entropy for emotional state both in 2D. In the eye state dataset 3D space is best, selected by Symmetrical Uncertainty. Finally, 10-fold cross validation is used to train best topologies. Final best 10-fold results are 97.03% for attention state (2D CNN), 98.4% for Emotional State (3D CNN), and 97.96% for Eye State (3D CNN). Significance: The findings of the framework presented by this work show that CNNs can successfully convolve useful features from a set of pre-computed statistical temporal features from raw EEG waves. The high performance of K-fold validated algorithms argue that the features learnt by the CNNs hold useful knowledge for classification in addition to the pre-computed features.