NTU Researchers Create an ML Model, Ycogni, Capable of Screening Smartwatch Data for Depression

Scientists from the Nanyang Technological University Singapore (NTU Singapore) have developed a machine learning system, dubbed Ycogni, which they say can process data from wearable devices to detect depression.

"Our study successfully showed that we could harness sensor data from wearables to aid in detecting the risk of developing depression in individuals," says study co-lead Josip Car, professor and director of the NTU Center for Population Health Sciences, of the work. "By tapping on our machine learning program, as well as the increasing popularity of wearable devices, it could one day be used for timely and unobtrusive depression screening."

The project saw groups of depressed and healthy participants being monitored using off-the-shelf Fitbit Charge 2 wearables for sleep and activity patterns based on step counting, heart rate, estimated energy expenditure, and sleep monitoring — the majority being features found in even the lowest-cost smart watches on the market today.

Fed into a machine learning system, the study was able to make conclusions about possible links between vital signs and depression - such as confirming earlier observations that subjects with varied heart rates in the early hours of the morning were more likely to experience severe depressive symptoms than others, and that irregular sleeping patterns had a similar indicative effect.

"This is a study that, we hope, can set up the basis for using wearable technology to help individuals, researchers mental health practitioners and policy makers to improve mental well-being," says Georgios Christopoulos, associate professor at NTU's Nanyang Business School. "But on a more generic and futuristic application, we believe that such signals could be integrated with Smart Buildings or even Smart Cities initiatives: Imagine a hospital or a military unit that could use these signals to identify people-at-risk."

"Our team will also be working on expanding to other types of psychological status, such as mental fatigue, which seems to be an alarming problem nowadays. Wearables can also be part of feedback system that could support therapists to better evaluate the psychological status of their patients — for instance improvements in sleep quality."

"We look forward to expanding on our research to include other vital signs in the detection of depression risk, such as skin temperature," Car adds. "Fine-tuning our program could help in facilitating early, unobtrusive, continuous, and cost-effective detection of depression in the general population."

The team's work has been published in the journal JMIR mHealth and uHealth, and is available under open-access terms from the NTU website as a PDF download.