Continuous Detection of Physiological Stress with Commodity Hardware.

Area of Science:

• Physiological computing and biosignal processing.
• Wearable technology and human-computer interaction.
• Stress detection and health monitoring.

Background:

• Effective stress management is crucial for preventing adverse health outcomes.
• Current accurate stress detection methods often use expensive, bulky clinical-grade sensors, limiting accessibility.
• There is a need for affordable, non-clinical solutions for continuous stress monitoring.

Purpose of the Study:

• To investigate the feasibility of using commercially available, off-the-shelf (COTS) sensors for stress monitoring.
• To develop and evaluate a system for inferring stress levels from physiological signals captured by COTS sensors.
• To compare the performance of COTS sensors against clinical-grade sensors in stress detection.

Main Methods:

• Development of a stress monitoring system utilizing a COTS heart rate monitor (Polar H7).
• Evaluation of the system with 26 participants in controlled laboratory settings using validated stress-inducing stimuli.
• Field testing of the system under free-living conditions to assess real-world performance.

Main Results:

• The system demonstrated significant accuracy in detecting stressful events using only the COTS heart rate monitor.
• Achieved an F1-score of up to 0.87 in laboratory settings and 0.66 in field conditions.
• Performance was comparable to that of clinical-grade sensors, indicating high viability of the approach.

Conclusions:

• Commercially available off-the-shelf sensors are viable for accurate stress detection.
• This approach offers a cost-effective and accessible alternative to clinical-grade sensors for stress monitoring.
• The findings support the potential for widespread adoption of COTS-based wearable systems for stress management.