Machine Learning for Anxiety Detection Using Biosignals: A Review

Diagnostics, Journal Year: 2022, Volume and Issue: 12(8), P. 1794 - 1794

Published: July 25, 2022

Anxiety disorder (AD) is a major mental health illness. However, due to the many symptoms and confounding factors associated with AD, it difficult diagnose, patients remain untreated for long time. Therefore, researchers have become increasingly interested in non-invasive biosignals, such as electroencephalography (EEG), electrocardiogram (ECG), electrodermal response (EDA), respiration (RSP). Applying machine learning these signals enables clinicians recognize patterns of anxiety differentiate sick patient from healthy one. Further, models multiple diverse biosignals been developed improve accuracy convenience. This paper reviews summarizes studies published 2012 2022 that applied different algorithms various biosignals. In doing so, offers perspectives on strengths weaknesses current developments guide future advancements detection. Specifically, this literature review reveals promising measurement accuracies ranging 55% 98% sample sizes 10 102 participants. On average, using only EEG seemed obtain best performance, but most accurate results were obtained EDA, RSP, heart rate. Random forest support vector machines found be widely used methods, they lead good feature selection has performed. Neural networks are also extensively provide accuracy, benefit no needed. comments effective combinations modalities success detecting anxiety.

Language: Английский

---

An Individual-Oriented Algorithm for Stress Detection in Wearable Sensor Measurements

IEEE Sensors Journal, Journal Year: 2023, Volume and Issue: 23(19), P. 22845 - 22856

Published: Aug. 16, 2023

Accurately measuring a person's level of stress can have wide variety impacts, not only for human health, but also the perceived feeling safety when going after daily habits, such as walking, cycling, or driving from one place to another. While there is vast amount research done on and related physiological responses body, no go-to method it comes acute in live setting. This work proposes an advancement rule-based detection algorithm proposed [1], identify moments (MOS) more reliably, through adaptation individualization rules original paper. The leverages electrodermal activity skin temperature, both recorded by Empatica E4 wristband, assessment individual's exposed audible stimulus. achieves average recall 81.31%, with precision 46.23%, accuracy 92.74%, measured 16 test subjects. trade-off between be controlled adjusting MOS threshold that needs reached detected.

Language: Английский

---

Stress-Wed: Stress Recognition Autoencoder Using Wearables Data

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 77 - 88

Published: Jan. 1, 2025

Language: Английский

---

Fusing Wearable Biosensors with Artificial Intelligence for Mental Health Monitoring: A Systematic Review

Biosensors, Journal Year: 2025, Volume and Issue: 15(4), P. 202 - 202

Published: March 21, 2025

The development of digital instruments for mental health monitoring using biosensor data from wearable devices can enable remote, longitudinal, and objective quantitative benchmarks. To survey developments trends in this field, we conducted a systematic review artificial intelligence (AI) models biosensors to predict conditions symptoms. Following PRISMA guidelines, identified 48 studies variety smartphone including heart rate, rate variability (HRV), electrodermal activity/galvanic skin response (EDA/GSR), proxies biosignals such as accelerometry, location, audio, usage metadata. We observed several technical methodological challenges across lack ecological validity, heterogeneity, small sample sizes, battery drainage issues. outline corresponding opportunities advancement the field AI-driven biosensing health.

Language: Английский

---

A multi-modal deep learning approach for stress detection using physiological signals: integrating time and frequency domain features

Frontiers in Physiology, Journal Year: 2025, Volume and Issue: 16

Published: April 1, 2025

This study aims to develop a multimodal deep learning-based stress detection method (MMFD-SD) using intermittently collected physiological signals from wearable devices, including accelerometer data, electrodermal activity (EDA), heart rate (HR), and skin temperature. Given the unique demands high-intensity work environment of nursing profession, measurement in nurses serves as representative case, reflecting levels other high-pressure occupations. We propose learning framework that integrates time-domain frequency-domain features for detection. To enhance model robustness generalization, data augmentation techniques such sliding window jittering are applied. Feature extraction includes statistical derived raw obtained via Fast Fourier Transform (FFT). A customized architecture employs convolutional neural networks (CNNs) process separately, followed by fully connected layers final classification. address class imbalance, Synthetic Minority Over-sampling Technique (SMOTE) is utilized. The trained evaluated on signal dataset with level labels. Experimental results demonstrate MMFD-SD achieves outstanding performance detection, an accuracy 91.00% F1-score 0.91. Compared traditional machine classifiers logistic regression, random forest, XGBoost, proposed significantly improves both robustness. Ablation studies reveal integration plays crucial role enhancing performance. Additionally, sensitivity analysis confirms model's stability adaptability across different hyperparameter settings. provides accurate robust approach integrating features. Designed occupational environments intermittent collection, it effectively addresses real-world monitoring challenges. Future research can explore fusion additional modalities, real-time improvements generalization its practical applicability.

Language: Английский

---

Machine Learning Techniques for Arousal Classification from Electrodermal Activity: A Systematic Review

Sensors, Journal Year: 2022, Volume and Issue: 22(22), P. 8886 - 8886

Published: Nov. 17, 2022

This article introduces a systematic review on arousal classification based electrodermal activity (EDA) and machine learning (ML). From first set of 284 articles searched for in six scientific databases, fifty-nine were finally selected according to various criteria established. The has made it possible analyse all the steps which EDA signals are subjected: acquisition, pre-processing, processing feature extraction. Finally, ML techniques applied features these have been studied. It found that support vector machines artificial neural networks stand out within supervised methods given their high-performance values. In contrast, shown unsupervised is not present detection through EDA. concludes use widely spread, with particularly good results found.

Language: Английский

---

Utilizing Machine Learning for Context-Aware Digital Biomarker of Stress in Older Adults

Information, Journal Year: 2024, Volume and Issue: 15(5), P. 274 - 274

Published: May 12, 2024

Identifying stress in older adults is a crucial field of research health and well-being. This allows us to take timely preventive measures that can help save lives. That why nonobtrusive way accurate precise detection necessary. Researchers have proposed many statistical measurements associate with sensor readings from digital biomarkers. With the recent progress Artificial Intelligence healthcare domain, application machine learning showing promising results detection. Still, viability for biomarkers under-explored. In this work, we first investigate performance supervised algorithm (Random Forest) manual feature engineering contextual information. The concentration salivary cortisol was used as golden standard here. Our framework categorizes into No Stress, Low High Stress by analyzing gathered wearable sensors. We also provide thorough knowledge combining physiological data obtained sensors clues protocol. context-aware model, using fusion, achieved macroaverage F-1 score 0.937 an accuracy 92.48% identifying three levels. further extend our work get rid burden engineering. explore Convolutional Neural Network (CNN)-based encoder detect in-depth look at CNN-based encoder, which effectively separates useful features inputs. Both frameworks, i.e., Random Forest engineered Fully Connected validate integration more insight response even without any self-reporting or caregiver labels. method fusion shows 83.7797% 0.7552, respectively, context 96.7525% 0.9745 context, constitutes 4% increase 0.04 RF.

Language: Английский

---

TAGformer: A Multimodal Physiological Signals Fusion Network for Pilot Stress Recognition

IEEE Sensors Journal, Journal Year: 2024, Volume and Issue: 24(13), P. 20842 - 20854

Published: May 14, 2024

Pilot stress recognition is crucial for safe and smooth flight, while heightened can significantly impede pilots' capacity to respond potential dangers. Recent research has witnessed the success of deep learning models using multimodal physiological signals in achieving high classification accuracy. However, these often overlook intricate dependencies among signals, especially as they vary with different levels. Explicitly modeling enhance feature extraction efficiency a more compact network model, thus improving accuracy recognition. Therefore, we propose novel model pilot based on 14 data, including electrocardiography (ECG), electromyography (EMG), heart rate (HR), respiration (RESP), skin temperature (SKT). Handcrafted features from data are initially organized into graph fused topology adaptive convolutional module (TAGCM). Then, extracted fed transformer encoder followed by multilayer perceptron (MLP) recognizing stress. The multistage gated average fusion (MGAF) was employed fuse modules. Experiments were conducted self-collected dataset well publicly available drivers' experimental results show that proposed could achieve better terms ability levels than other baseline methods. Moreover, outcomes obtained experiment public underscore effectively across diverse scenarios.

Language: Английский

---

Automated Detection of Mental Stress Using Multimodal Characterization of PPG Signal for AI Based Healthcare Applications

SN Computer Science, Journal Year: 2024, Volume and Issue: 5(6)

Published: July 29, 2024

Language: Английский

---

Real-Time Stress Detection from Raw Noisy PPG Signals Using LSTM Model Leveraging TinyML

Arabian Journal for Science and Engineering, Journal Year: 2024, Volume and Issue: unknown

Published: May 7, 2024

Language: Английский

---