Green Door Clinic, a specialized mental health clinic in the Harley Street Medical Area (HSMA), is testing a machine learning (ML) tool that detects voice changes in a novel manner to treat depression.

One issue with treating serious depression is that patients and therapists frequently lack a reliable method of tracking the illness’s course over time. Patients may struggle to remember or monitor how they have been feeling between visits if there are long gaps between them. Information can be overlooked, which is exacerbated if a patient sees multiple clinicians.

Jenny Dew, a lead psychotherapist at Green Door Clinic, said, “The director of Green Door clinic Rami Alfalahi has an extensive background in computers, working for several international companies for many years as a technology expert. Technology has always been a passion for him, so having the chance to help combine the worlds of mental health and computer science was an exciting and inspiring opportunity.

“The Green Door Clinic helps provide both patients and therapists to assist in the trialling of the technology. We do this because we believe in innovation and technological improvement in the health industry. Advances in technology to improve the treatment of mental health has been an area without much development for many years. Therefore, when informed of the project by Imperial College, we jumped at the opportunity to assist.”

Towards resolving this challenge, Affect.AI was developed as a voice-based AI tool. Dr. Wun Wong along with Imperial College students Woochan Hwang and Alice Tang developed a system that uses machine learning along with clinical assessments to determine if the characteristics of someone’s voice change over time so that a clinician can monitor changes in that person’s mood. In addition to being used between appointments, this tool provides clinicians with a means to evaluate treatment regimens based on voice recordings.

Patients undergoing depression treatment were the first to test the technology in 2019/2020. Intensive therapy is being tested in the UK at Green Door Clinic, which specializes in providing it as an alternative to standard weekly therapy sessions.

In addition to identifying the symptoms of depression, vocal traits such as pitch and volume can be used as biomarkers to determine the severity of the illness. The rise of advanced computer analytics-enabled researchers to quantify these paraverbal patterns after studying them for over a decade. By comparing the data collected over time with the clinician’s own baseline, computer algorithms can detect subtle differences in tone that a human might miss.

Their treatment plans are based on the latest research that Green Door Clinic uses to provide personalized mental healthcare. The clinic uses a wide variety of therapeutic protocols to maximize patients’ recovery and stay abreast of the latest developments in neuroscience.