A systematic review of digital and wearable sensor data in the ABCD study: Insights, methods, and emerging...

Research into key adolescent health domains, like problematic screen-based media use, sleep health, and physical activity, are frequently limited by error-prone self-report-based measures. To improve the measurement of such behaviors, the Adolescent Brain Cognitive Development Study (ABCD), a diverse longitudinal cohort study of nearly 12,000 U.S. youth (ages 9-10 at Baseline), includes remote sensing protocols during data collection. Consenting participants provided up to three weeks of Fitbit actigraphy data, beginning at the year-two follow up, and smartphone usage data, beginning at the year-four follow up, collected via the Effortless Assessment of Research Systems (EARS) smartphone application. This systematic review aimed to identify and synthesize all published empirical studies which have utilized the ABCD Study's passive sensor data, identifying 45 articles investigating measurement validity, and associations with adolescent health, development, and neurocognitive functioning. Fitbit devices broadly reproduced estimates of sleep and physical activity obtained from research-grade sensors. Smartphone sensor data were internally consistent, reproduced expected daily usage patterns, and were moderately correlated with self-report. Most substantive studies utilized Fitbit data and reported cross-sectional associations between sensor-derived single-item summary statistics and survey-based measures of mental and behavioral health, neurocognitive functioning, or environmental risk factors. A smaller number of studies utilized longitudinal designs, within-person analyses, machine learning approaches, and/or genetic or neuroimaging methods to reduce confounding and enhance biological and etiological insight into these observed relationships. While published findings were largely incremental in nature, ABCD's passive sensing resources nonetheless represent a powerful, widely available resource for incorporating sensor-based measures into adolescent health research.