KAIST: "Reading Mental Health Through Internet of Things Data"

Research has shown that by utilizing Internet of Things (IoT) data, it is possible to precisely track a user's mental health status and detect changes in daily rhythms-signals that are difficult to capture with only smartphones or wearable devices-enabling the early identification of deteriorating mental health. This finding is expected to lay the foundation for the development of personalized mental health management systems in the future.

On October 21, KAIST announced that a research team led by Professor Euijin Lee in the Department of Computer Science had demonstrated the potential to precisely track an individual's mental health status using IoT sensor data collected within the home.

(Photo) From the left, Chanhee Lee, PhD candidate in the Department of Computer Science, Euijin Lee, Professor in the Department of Computer Science, Hyunsoo Lee, Professor in the Department of Computer Science, and Youngji Ko, PhD candidate in the Department of Computer Science. Provided by KAIST

Recently, the number of single-person households in South Korea has surpassed 8 million, accounting for 36% of all households-a record high. However, the worsening of mental health due to factors such as loneliness and isolation among single-person households is emerging as a social issue. In fact, the Seoul Metropolitan Government recently reported that 62% of all single-person households in the city are experiencing mental health issues such as loneliness.

Consistently monitoring one's own condition is essential for effective mental health management. With this in mind, the research team focused on environmental data from within the home. However, conventional smartphone- and wearable-based tracking methods have limitations, as it is difficult to collect data inside the home when users are not wearing or carrying these devices.

To address these limitations, the research team utilized home IoT sensors capable of continuously measuring daily activities without requiring any additional user intervention. Unlike previous studies that focused on disease diagnosis, this approach enables IoT users to independently track and proactively manage their emotional well-being-such as depression, anxiety, and stress-experienced in everyday life at home.

The team also conducted a four-week empirical study involving 20 single-person households of young adults. The study involved installing home appliances, sleep mats, and motion sensors to collect lifestyle data, which was then compared and analyzed against participants' self-reported mental health statuses.

The research team emphasized that the accuracy of models predicting risks of depression, anxiety, and stress improved significantly when IoT data was incorporated, compared to using only smartphone and wearable data.

In particular, the empirical study found a clear association between reduced sleep duration and increased indoor temperatures with worsening mental health. Additionally, participants exhibited distinct behavioral patterns in response to stress, such as an "overeating type" characterized by increased refrigerator use, and a "lethargic type" marked by a sharp decline in activity. At the same time, the more irregular the participants' daily routines, the more likely their mental health was to deteriorate-a common trend observed across individuals.

This suggests that the variability in daily routines is a more significant factor than the frequency of specific behaviors, highlighting that maintaining regular daily patterns is key to sustaining mental health.

After completing the study, participants reported that when they reviewed their lifestyle data through the visualized software, they found the data to be genuinely helpful for understanding their mental health, rather than being concerned about privacy invasion. The research team noted that this demonstrates high acceptance of the research and a high level of participant satisfaction.

Professor Lee stated, "This study demonstrates that in-home IoT data can serve as an important clue for understanding mental health within the context of an individual's daily life. In the future, we expect this to contribute to the development of remote healthcare systems that utilize artificial intelligence (AI) to predict individual lifestyle patterns and provide personalized coaching."

This research was conducted with support from the LG Electronics-KAIST Digital Healthcare Research Center and the National Research Foundation of Korea, funded by the Ministry of Science and ICT.

This content was produced with the assistance of AI translation services.