A technology has been developed that can predict tomorrow's mood like a weather forecast using a wearable device, and it is expected to help treat patients with mood disorders such as depression and mania.

The CI research team led by Jae-kyung Kim of the Group for Mathematical and Computational Sciences at the Institute for Basic Science (IBS) announced on the 25th that they have developed a technology to predict mood based on sleep patterns together with the team of Professor Heon-jeong Lee of Korea University.

The researchers developed a model that predicts mood episodes using only sleep-wake pattern data.

Sleep data were collected for an average of 429 days from 168 mood disorder patients.

The researchers applied this data to a machine learning algorithm and were able to predict the next day's depression and mania with 80% and 98% accuracy, respectively.

We found that changes in circadian rhythms are key predictors of mood episodes. The risk of depressive episodes increases as the circadian rhythm becomes delayed.

Conversely, if it is brought forward excessively, the risk of manic episodes increases.

For example, people who have a rhythm of going to bed at 11 p.m. and waking up at 7 a.m. are at higher risk of developing depressive episodes if they go to bed late and wake up late.

The methodology presented by the research team is expected to increase the treatment effectiveness of patients with mood disorders.

In actual clinical practice, light therapy is performed in the early morning to treat patients with seasonal depression.

This study presents a methodology for obtaining objective mood episode data.

A particular advantage is that data can be acquired non-invasively and passively through wearable devices.

Professor Lee Heon-jeong said, “This study presents a new paradigm for predicting mood disorders.”

“We reduced the cost of data collection and increased the possibility of clinical application,” said Kim Jae-kyung, CI.

The results of this study were published online in 'NPJ Digital Medicine' on November 18.