Scientists use video games to detect depression

Thousands of patients are taking part in a pilot to test new artificial intelligence (AI) software that analyses data from video games to assess whether the user is depressed.

Thymia, a London-based start-up founded by neuroscientist Emilia Molimpakis and theoretical physicist Dr Stefano Goria, has created a software platform that requires patients to play specially-designed video games. These games include verbally describing animated scenes and memorising moving objects such as bees.

While users are playing the games, the Thymia software analyses three key data streams: voice, video and behavioural measures. It then identifies data patterns that might indicate depression, enabling the software to suggest a diagnosis. Analysis of what someone is saying and how they are saying it can pick up clues as to their state of mind. Analysis of the video can pick up micro-expressions and eye gaze, helping to track mood, while behavioural measures (such as reaction times, memory, and error rates) can help detect how severe the depression is.

It will also show if any treatments are working over a period of time, so that clinics can identify the appropriate combination of treatments for a patient.

The platform will enable clinicians to monitor patients remotely at home in the weeks between in-person appointments, helping doctors and patients build an in-depth understanding of their condition over time.

Too many cases of depression are undiagnosed

Currently depression is diagnosed largely by administering questionnaires to patients. The aim of the Thymia software is to remove the subjective element from diagnosis. Molimpakis, the company’s CEO, said: “GPs and psychiatrists are using the same diagnostic questionnaires that have been around since the 1960’s. Rating how sad you feel on a scale of 0 to 3 simply is not enough to capture the subtle nuances of early signs of depression nor track the complexities of ongoing mental illness, but our clinicians have not been given a better option. This means too many cases go undiagnosed or are misdiagnosed, and too many patients wait years before the right treatment is found. We want to empower clinicians and patients themselves with better tools.”

The software is now being trialled by patients in London, including at UCL and King’s College London, to train the AI algorithm and finetune the user experience.

“Thymia is bringing psychiatry into the 21st Century, with an AI-enabled platform for accurate and continuous patient monitoring,” said Molimpakis. “It’s the first objective psychiatric assessment system, combining multiple layers of rich physiological data to assess depression and distinguish between similarly presenting disorders.”

She added: “Our technology will empower clinicians to assess and treat depression sooner, whilst allowing patients to develop a deeper understanding of their own condition.

“In time our aim is to become the gold standard of assessment for all mental health disorders and show that mental illness is as real and objectively measurable as physical illness, thereby also helping eradicate the stigma associated with it.”