Neuro into the network: a way has been found to more accurately diagnose schizophrenia using MRI

Scientists have developed a method for analyzing brain activity that allows you to assess disorders in its work at different levels. The approach is based on functional magnetic resonance imaging at rest. The researchers also comprehensively described for the first time how disruptions of neural connections in the brain at different levels affect the development of the disorder. The developed technique, combined with artificial intelligence approaches, opens the way to more accurate diagnosis and personalized therapy of schizophrenia, experts say. For more information about the study, see the Izvestia article.

A new diagnostic method for schizophrenia

Scientists from the Plekhanov Russian University of Economics and Samara State Medical University, together with colleagues from the Plovdiv Medical University (Bulgaria), have developed a multi-level method for analyzing brain activity, which allows assessing disorders in its work at different levels. The approach is based on functional magnetic resonance imaging (fMRI) at rest, a method of measuring brain activity at times when a person is not doing anything.

As scientists told Izvestia, research shows that the disease develops as a result of disrupted connections between nerve cells (neurons) in the brain. Thus, disruptions in the collaboration of large-scale networks (brain regions) lead to characteristic symptoms: hallucinations, delusions, impaired thinking and decreased motivation. However, it was not fully understood how the development of schizophrenia is affected by disorders at different levels of brain organization — from the interaction of brain regions to the connection between individual regions within these departments.

The researchers obtained fMRI data for 43 patients diagnosed with schizophrenia and 63 healthy volunteers from the control group. The authors analyzed the results on two different scales: They assessed the overall organization of neural networks of the entire brain (global level) and studied the functions of 15 large-scale networks, such as passive brain mode networks, visual and auditory networks, and others (macro level).

The authors of the study are Semyon Kurkin and Alexander Khramov.

Photo: Semyon Kurkin

The scientists then identified specific disrupted connections between individual brain regions using network statistics, a mathematical method used to identify significant network characteristics in the brain. To visually present the results, the authors applied an original multigraph approach that allows for different types of connections and shows how interactions between neural networks change in a patient with schizophrenia compared with a healthy brain.

IT with a nose: scientists create "therapeutic perfumes" for fear of doctors

How AI can Help you find an anti-anxiety scent

— The proposed method serves as a powerful and versatile tool for analyzing MRI data and can be applied to other neuropsychiatric diseases such as bipolar disorder and depression. Identifying the links between changes in neural networks and the main symptoms opens up opportunities for the development of more targeted treatment and diagnostic methods," said Semyon Kurkin, Doctor of Physico—Mathematical Sciences, Chief Researcher at the Research Institute of Applied Artificial Intelligence and Digital Solutions at the Plekhanov Russian University of Economics.

In the future, scientists plan to expand the scope of the new method using the example of other neuropsychiatric and neurological diseases, as well as combine it with artificial intelligence algorithms to develop promising diagnostic systems.

Causes of schizophrenia

The study showed that at the global level, the neural networks of patients with schizophrenia have an increased density of connections compared to the norm, in addition, these networks tend to form local groups. As a result, there are more connections between brain regions, but the efficiency of these regions decreases as they begin to interfere with each other, creating "information noise." Because of this, the brain spends more resources on useless work.

Functional magnetic resonance imaging of a volunteer participating in the experiment

Photo: Semyon Kurkin

The researchers also found a weakening of the connection between the temporal (which reads signals from the whole body), orbitofrontal (responsible for social behavior) and cingulate (directing the focus of attention to important tasks) areas of the brain. This means that departments that must work together to filter information and manage attention do not communicate well with each other, the researchers explained. Thus, the authors found direct confirmation that a violation of the temporo-frontal lobe plays a central role in the development of symptoms of schizophrenia.

Memorable story: scientists have found a way to overcome "dementia" in AI

How the new technology will help create continuously learning neural networks

The study represents a significant methodological breakthrough due to the introduction of a comprehensive multilevel analysis of functional connectivity (the ability to interact), which for the first time systematically described the hierarchy of network disorders in schizophrenia, Alexander Zakharov, director of the Research Institute of Neuroscience at SamSMU, an expert at the NTI Center based at SamSMU and the NTI Helsnet market, told Izvestia.

—An innovative approach integrating the assessment of global topology, the interaction of large-scale networks and local circuits has allowed us to resolve the apparent paradox of the simultaneous existence of signs of hypo- and hyperconnectivity," he noted.

Project leader Semyon Kurkin and graduate student Nikita Kulagin in the process of discussing the results

Photo: Semyon Kurkin

From a practical point of view, specific network biomarkers open the way to the development of more accurate diagnostic tools. In the future, they may form the basis of neuroimaging tests for objective differential diagnosis and evaluation of the effectiveness of therapy, Alexander Zakharov concluded.

Research in this area has been conducted for a long time, noted NTI expert, neurosurgeon Artur Biktimirov.

"We can talk about the increment of new scientific data for further breakthroughs in science and, subsequently, in technology," he believes.

The results of the study, supported by a grant from the Russian Science Foundation (RSF), are published in the journal Psychiatry Research: Neuroimaging.