Researchers at Washington University have developed a promising deep-learning model called WearNet that utilises wearable technology to detect mental disorders. By analysing 10 variables collected by Fitbit activity trackers, such as daily steps, calorie burn rates, and heart rate, WearNet surpassed state-of-the-art machine learning models in identifying depression and anxiety. What sets WearNet apart is its ability to provide individual-level predictions of mental health outcomes, unlike traditional group-level statistical analyses. This groundbreaking study involved over 10,000 Fitbit users, making it the largest and most diverse cohort examined thus far.The researchers highlight the significance of early detection in mental disorder management and recognize the potential of wearables in facilitating diagnosis and treatment. Wearable data offers an unobtrusive approach to identifying these disorders, allowing clinicians to intervene promptly and individuals to make behavioural adjustments. Previous studies often relied on small cohorts, whereas this study benefited from data sourced from the National Institutes of Health's "All of Us" research program, providing comprehensive demographic information.
Presented at the ACM/IEEE Conference on Internet of Things Design and Implementation, the Washington University team's findings received the Best Paper Award for IoT Data Analytics. The researchers envision their AI model as an automated screening tool that can inform individuals about their mental health status and recommend professional assistance. Ultimately, their goal is to persuade hospitals or companies to implement this technology and meet the pressing need for effective and non-invasive methods of detecting mental disorders.