Personalized Stress Detection Using Biosignals from Wearables: A Scoping Review.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence
• Wearable Technology

Background:

• Stress is a pervasive aspect of modern life requiring effective management strategies.
• Wearable technology offers a promising avenue for continuous, personalized stress monitoring.
• Existing research on personalized stress detection models needs systematic synthesis.

Purpose of the Study:

• To conduct a scoping review of personalized stress detection models utilizing wearable technology.
• To identify key biosignals, AI methodologies, datasets, and implementation challenges.
• To synthesize current knowledge and identify gaps for future research.

Main Methods:

• Systematic literature search using PRISMA-ScR framework across Scopus, IEEE Xplore, and PubMed.
• Analysis focused on biosignals (EDA, PPG), AI techniques (deep learning), datasets, wearable devices, and practical challenges.
• Synthesis of findings on stress detection model components and deployment issues.

Main Results:

• Electrodermal activity (EDA) and photoplethysmography (PPG) are prominent biosignals for stress detection.
• A trend towards deep learning models in stress detection was observed, though comparisons with traditional methods are limited.
• Challenges include dataset representativeness, data quality, privacy concerns, and real-world deployment hurdles.

Conclusions:

• Multimodal biosignal analysis, particularly using EDA and PPG, shows potential for reliable stress detection.
• Further research is essential to validate AI models against traditional methods and address practical implementation barriers.
• Future work should focus on creating comprehensive datasets and developing user-centric, efficient AI for real-world stress management systems.