This app can predict moods

Disu Dang

Depression, a global health burden affecting millions, remains a significant challenge in Hong Kong. With an estimated prevalence of 8.3 percent according to the 2015 Hong Kong Mental Morbidity Survey, the impact of depression on the population is undeniable.

However, the survey also revealed that a significant majority - about 70 percent - of individuals experiencing common mental disorders such as depression and anxiety, did not seek help from mental health services.

This lack of utilization can be attributed to insufficient awareness, societal stigma, and limited accessibility to these services.

In response to this pressing issue, the Faculty of Medicine at the Chinese University of Hong Kong has developed an innovative mobile depression assessment app that uses multimodal artificial intelligence analysis.

By using AI to analyze users' facial expressions, voice, language, subjective mood state, and rest-activity statistics, the app has demonstrated an accuracy of 81 percent in predicting major depressive disorder.

Wing Yun-kwok, Choh-Ming Li professor of psychiatry and chairman of the Department of Psychiatry at CU Medicine, explained: "Depression is more than simply sadness but encompasses a series of physiological, cognitive, mood, language and rest-activity changes.

"Therefore, other than conventional assessment by clinicians, digital phenotyping, which measures a series of parameters, has great potential to help with the assessment and monitoring of depression."

The app incorporates a unique approach to analyzing users' mental status. It prompts users to input a mood diary four times a day for a consecutive seven-day period. The mood diary involves recording a video where users share their current activities and emotions, and self-rating the users' mood on a scale of one to 10. It takes about two minutes for the entire process.

Watson Chen Jie, a postdoctoral fellow in the Department of Psychiatry at CU Medicine, highlights several key indicators employed in the app's analysis.

Firstly, the app uses AI to analyze users' facial expressions, focusing on specific patterns observed among MDD subjects, such as more eyebrow narrowing and less lip corner pulling. It also examines users' speech patterns, identifying linguistic tendencies.

Individuals with MDD tend to refer to themselves more frequently during conversations and use more negative emotion words. Additionally, the app detects variations in articulation rate and pause duration, with MDD subjects being slower in speech and using longer pauses. In addition, the app incorporates the use of an actiwatch to monitor users' rest-activity patterns.

The results showed that MDD subjects exhibited reduced mobility, spending less time engaged in physical activity, and experiencing delayed sleep midpoint.

"By combining the conventional self-reporting scale with multimodal assessment, performance in predicting the non-remission status of MDD can reach an F1 score of 0.70," said Chen.

Tim Li Man-ho, assistant professor in the Department of Psychiatry, said that the app is currently used solely for research purposes involving depression patients. To expand its application for assessing depression in the broader community, the research team recognizes the need for a larger and more diverse sample size.

The team's plan to develop a real-time feedback system, which could function as a virtual therapist.

"Depression is very common and the demand for its health services is huge. The new technology has the potential to assist in diagnosis and monitoring of this health issue, helping to alleviate the pressure on healthcare providers," said Li.

disu.dang@singtaonewscorp.com