Smartwatches that track physical and physiological data could transform how we understand and treat brain diseases and behavioral disorders, according to a study published in Cell. Led by Mark Gerstein of Yale University and involving Diego Garrido Martín from the University of Barcelona, the research used data from over 5,000 adolescents to explore the connection between wearable sensor data, psychiatric illnesses, and genetics.
The study utilized smartwatch data collected from adolescents aged 9–14 as part of the Adolescent Brain Cognitive Development Study, the largest U.S. project on brain development. The data included heart rate, calorie burn, physical activity, step count, and sleep patterns. Researchers processed this information to create what they termed a “digital phenotype,” traits measured and tracked using digital tools.
Artificial intelligence (AI) models trained on this data could predict psychiatric illnesses like ADHD and anxiety. Heart rate proved crucial for identifying ADHD, while sleep quality and stages were key for recognizing anxiety. The study also showed that wearable data could help distinguish between subtypes of these disorders, such as inattention versus hyperactivity within ADHD, which require different treatments.
Beyond diagnosis, the study explored links between digital phenotypes and genetics. By analyzing smartwatch data and genetic information together, researchers identified 37 genes associated with ADHD. However, they found no direct genetic links to an ADHD diagnosis without the smartwatch data, highlighting the added value of continuous, detailed physiological tracking.
The methodology involved converting raw smartwatch data into formats usable by AI, a technically challenging process. This innovation enables a more nuanced understanding of psychiatric illnesses and their genetic foundations, bridging the gap between behavior and biology.
The findings suggest that wearable technology could revolutionize psychiatry by offering precise, quantitative tools for diagnosis and treatment. The approach may extend beyond psychiatric conditions to other neurological diseases and disorders, offering a broader application of digital phenotyping.
Co-author Walter Roberts of Yale emphasized that this method could reshape psychiatry by integrating behavioral data with genetic insights. While the current focus is on ADHD and anxiety, the researchers believe this approach holds promise for studying neurodegeneration and other complex conditions. Ultimately, these advancements may inspire a shift away from traditional clinical diagnostics toward more data-driven, personalized healthcare solutions.
- Source: Press release University of Barcelona
