Identifying Biomarkers for Accurate Detection of Stress

Area of Science:

• Biomedical Engineering
• Neuroscience
• Data Science

Background:

• Substance use disorder (SUD) is a growing epidemic with severe neurological and behavioral consequences.
• Stress is a significant risk factor for SUD development, relapse, and drug seeking.
• Current SUD treatments are limited, necessitating novel detection and management strategies.

Purpose of the Study:

• To investigate physiological stress markers for SUD.
• To identify optimal biomarkers for stress detection using machine learning.
• To differentiate between mental and physical stress and analyze meditation's effects.

Main Methods:

• Collected eight physiological features from fifteen individuals using a RespiBAN Professional chest-worn device.
• Selected optimal biomarkers for stress detection via machine learning classifiers.
• Analyzed differences between mental and physical stress and the impact of meditation.

Main Results:

• Electrodermal activity (EDA), body temperature, and chest accelerometer data were identified as the most accurate biomarkers for stress detection.
• Machine learning classifiers effectively distinguished stress based on these features.
• Distinct patterns were observed for mental versus physical stress and during meditation.

Conclusions:

• EDA, body temperature, and accelerometer data are promising objective measures for stress detection in SUD.
• These findings support the development of new wearable devices for SUD management.
• Future research will focus on validating these methods in individuals with SUD.