A new smartphone application called MoodCapture is set to transform mental health care by using AI to detect early signs of depression through facial cues. The app utilizes a phone’s front camera to analyze users’ facial expressions and surroundings, offering real-time digital support for mental health.
Developed by researchers from Dartmouth’s Department of Computer Science and Geisel School of Medicine, MoodCapture promises to reliably identify symptoms of depression with a remarkable 75% accuracy rate. This innovation is particularly significant given its passive, nonintrusive methodology, leveraging natural facial expressions captured during regular phone use.
The app’s development is the culmination of extensive research, including a study involving 177 individuals diagnosed with major depressive disorder. Through capturing and analyzing 125,000 images over 90 days, MoodCapture’s AI was trained to detect depressive symptoms based on specific facial expressions and environmental factors, such as lighting and the presence of others.
Andrew Campbell, the Albert Bradley 1915 Third Century Professor of Computer Science and the paper’s corresponding author, highlights the app’s unique approach to leveraging everyday interactions with technology. “This is the first time that natural ‘in-the-wild’ images have been used to predict depression,” Campbell said, emphasizing the potential for this technology to scale up seamlessly into daily life without additional user effort.
Nicholas Jacobson, a co-author of the study and assistant professor of biomedical data science and psychiatry, discussed the importance of capturing the fluctuating nature of depression, which traditional assessments often miss. “Our goal is to capture the changes in symptoms that people with depression experience in their daily lives,” Jacobson explained. This approach could significantly enhance the timeliness and effectiveness of interventions for depression.
The researchers foresee the technology becoming publicly available within the next five years, pending further development and refinement. The study’s results were published on the arXiv preprint database in advance of their presentation at the Association of Computing Machinery’s CHI 2024 conference in May.
In addition to its diagnostic capabilities, the researchers are focused on improving MoodCapture’s privacy features. Future iterations of the app might process images directly on the user’s device, ensuring data privacy while still benefiting from the AI’s analytical power.
