South Korean researchers have developed an AI-based tool that can predict episodes of mood disorder in patients using only sleep-wake data recorded by wearable devices such as smartwatches. People suffering from mood disorders, including bipolar disorder, experience long periods of sadness, depression, joy or mania. Mood disorders are closely linked with sleep-wake rhythms with disturbances potentially triggering a mood episode.
‍
The team comprised researchers from the Institute for Basic Sciences (IBS), Korea Advanced Institute of Science and Technology (KAIST), and Korea University College of Medicine.
‍
First, 168 Korean patients with mood disorders, including major depression and bipolar disorder, had their Fitbit sleep-wake data gathered and analyzed for 429 days. It is said that irregular sleep and circadian rhythms are intimately linked to mood problems, reports MobiHealthNews.
‍
In order to train models based on the machine learning package XGBoost to predict mood episodes, scientists took 36 sleep and circadian rhythm parameters out of this dataset.
‍
According to the results, the predictive models were able to predict depressive, manic, and hypomanic episodes with 80%, 98%, and 95% accuracy, respectively.
‍
Related Researchers Deploy Machine Learning to Improve Mental Health
‍
Findings also suggested that daily changes in circadian rhythm may be a key predictor of mood episodes; delayed circadian rhythms can potentially lead to depressive episodes while advanced circadian rhythms can raise the possibility of manic episodes.

‍
Although AI models for forecasting the emotional states of individuals with mood disorders already exist, they usually require a variety of data, such as GPS, sleep, heart rate, light exposure, and phone usage, all of which can be expensive to gather. The research team emphasized that it might also provide storage issues as well as security and privacy problems.
‍
"By developing a model that predicts mood episodes based solely on sleep-wake pattern data, we have reduced the cost of data collection and significantly improved clinical applicability," claimed Kim Jae Kyoung, KAIST associate professor and IBS chief investigator.
‍
The research team also highlights the possibility of using digital treatments in conjunction with mood state prediction to assess the daily risk of relapsing mood episodes and encourage healthy circadian rhythms and sleep-wake cycles by sending out reminders.
‍
The study was published in Nature's npj Digital Medicine journal.