Predicting Positive Psychological States using Machine Learning and Digital Biomarkers from Everyday Wearable Data

This dataset comprises physiological and psychological data collected from 34 participants over an 8-day study period, designed to investigate the prediction of positive psychological states using wearable sensor data. Physiological data was continuously recorded using research-grade Empatica Embrace Plus smartwatches, which are FDA-cleared medical devices equipped with five sensors: optical photoplethysmogram (PPG), electrodermal activity (EDA), 3-axis accelerometer, gyroscope, and peripheral skin temperature. The continuous monitoring resulted in 6,528 hours of raw physiological data. Concurrent psychological assessments were conducted using Ecological Momentary Assessment (EMA), generating 247 observations with 4,446 self-reported labels across 18 distinct psychological states covering positive affect, negative affect, self-esteem, sense of meaning in life, and personal relationships. Data preprocessing procedures are detailed in the accompanying manuscript. This dataset is provided exclusively for research purposes. Ethics approval and informed consent details are presented in the manuscript.