Wearable devices for anxiety assessment: a systematic review

Area of Science:

• Biomedical Engineering
• Digital Health
• Psychophysiology

Background:

• Growing interest in wearable biosensors for anxiety disorder assessment.
• Electrocardiography (ECG) is a primary signal for cardiovascular monitoring.
• Respiratory, electrodermal activity (EDA), and photoplethysmography (PPG) also show potential.

Purpose of the Study:

• To systematically review and evaluate biosignals from wearable devices for anxiety detection.
• To compare the efficacy of individual versus combined signal approaches.
• To identify common features and classification methods used in anxiety detection.

Main Methods:

• Systematic review of 26 studies (2014-2024) using wearable device data for anxiety detection.
• Extraction of study design, signal types, features, classification methods, and accuracy.
• Calculation of pooled accuracies for single-signal and multi-signal strategies.

Main Results:

• Multi-signal approaches yielded higher pooled accuracy (81.94%) than single-signal approaches (76.85%).
• Electrocardiography (ECG) demonstrated the highest pooled accuracy (80.34%) as an individual signal.
• Common features included heart rate and heart rate variability (ECG), inspiratory-to-expiratory ratio (respiratory), skin conductance (EDA), and heart rate (PPG); Support Vector Machine was the main classifier.

Conclusions:

• Wearable devices possess significant clinical potential for anxiety detection.
• Multimodal biosignal approaches are more effective than single-signal methods.
• Future research should enhance algorithms, increase sample sizes, and validate in diverse settings.