According to the World Health Organization, depression affects about 280 million people worldwide every year. This condition often increases the risk of suicide. Correct diagnosis and timely treatment may prevent such risks. To this end, a young scholar at the Institute of Applied Mathematics and Computer Science of TSU (IAMCS TSU), Neda Firoz, is developing a new approach to the diagnosis of depression using AI. The neural network is being taught to identify the symptoms based on the analysis of audio, video, text material and genetic data of a patient.

– Now I am collecting data for AI training and I am about to test the neural networks with different architectures to choose the best option, says Neda Firoz, a PhD student at IAMCS TSU. – We will automatically identify speech patterns that indicate the presence of a depressive disorder in a person. We use audio, video and text recordings of conversations made during a medical appointment for the analysis. The neural network can identify stable speech combinations typical for depression, even if it does not sound like a real complaint.

The second type of information that the neural network will analyze is electroencephalogram (EEG) data. They have only recently been used to diagnose depression, and the EEG markers of this disease have already been revealed. After training, the artificial intelligence will be able to identify them automatically.

The third block of data contains the results of full genetic sequencing. While previously it was quite expensive, now with technological advancements the cost of sequencing has decreased hundred times. More and more people seek for this service. Some want to find out potential health risks, others try to analyze their past and learn as much as possible about themselves.

– The presence of genetic risks alone does not mean that a person is likely to have a particular pathology, says Sergei Aksyonov, research supervisor and associate professor at the Department of Theoretical Foundations of Informatics, IAMCS TSU. –However, in aggregate, the analysis of three types of information may significantly improve the diagnosis and make it more accurate. Most certainly the AI will not replace a doctor, but it can become a good auxiliary tool.

Moreover, the Institute of Applied Mathematics and Computer Science of TSU is one of the leaders in Russia in terms of AI application and development of new tools based on it.

At the same time, the Institute acts as a center of competence for other Russian universities. For example, experts from Russian universities receive advanced training in AI at TSU. Besides, TSU helps universities to develop the study programs in the field of artificial intelligence and supports their successful implementation.