Using Wearable Passive Sensing to Predict Binge Eating in Response to Negative Affect Among Individuals With...

Area of Science:

• Psychiatry
• Digital Health
• Biomedical Engineering

Background:

• Binge eating (BE) is a public health crisis often preceded by negative affect.
• Ecological momentary assessment (EMA) is limited in capturing real-time affect and physiological arousal.
• Wearable sensors offer continuous, objective measurement of physiological arousal, potentially improving BE prediction.

Purpose of the Study:

• To determine if wearable sensor data can distinguish between positive and negative affect states.
• To assess if machine learning algorithms using sensor data and EMA improve BE prediction compared to EMA alone.

Main Methods:

• Recruit 30 individuals with BE for a 4-week study.
• Utilize Fitbit Sense 2 wristbands for continuous heart rate and electrodermal activity monitoring.
• Develop machine learning algorithms to analyze sensor data and EMA reports for affect and BE prediction.

Main Results:

• Anticipated results will demonstrate the capability of sensor features to differentiate affect states.
• The study expects to show improved accuracy in predicting BE using a combined sensor and EMA approach.

Conclusions:

• Integrating wearable sensor data with EMA can offer new insights into the affect-BE relationship.
• Findings may pave the way for developing advanced digital interventions for binge eating disorder.