Researchwearablesbipolar disorderecological momentary assessmentdigital biomarkers
Wearable Data Undermines Weekly Mood Labels
7.1
A letter by Priyanka Renita D'Souza et al., published March 12, 2026 in JMIR Medical Informatics, critiques Wu et al.'s study using wearable data and machine learning to predict mood symptoms in bipolar disorder. The authors argue weekly backfilled mood labels risk recall bias and misalign with daily physiological signals, and they recommend ecological momentary assessment and daily stress measures to improve model validity and clinical reliability.
Key Points
- 1Criticize weekly backfilled mood labels that may misalign with daily wearable-derived physiological signals.
- 2Note physiological markers like heart rate and sleep are nonspecific and confounded by external stressors.
- 3Recommend ecological momentary assessment and daily stress measures to improve label accuracy and model validity.
Scoring Rationale
Practical methodological critique with actionable recommendations, limited by non-peer-reviewed letter and no new empirical data.
Sources
Public references used for this report.
Practice with real Ride-Hailing data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
250 free problems · No credit card
See all Ride-Hailing problems