<p>This article presents results from analysis of the most informative regions of the brain for the diagnosis of schizophrenia based on resting-state functional magnetic resonance imaging (fMRI) data using the previously developed method of functionally homogeneous regions (FHR) and the CONN functional atlas. The analysis was performed on data from 32 subjects diagnosed with schizophrenia and 36 people from the control group, captured using a Siemens tomograph. Verification was carried out using data from 19 patients and 29 subjects from the control group, captured on a General Electric tomograph. The eight most informative regions were identified. Analysis of these regions showed that a change in the composition of the training group significantly affected the list of the most significant regions. At the same time, analysis of these most significant regions for repeatability by varying the composition of subjects showed that of the eight regions identified as the most significant, four had repeatability of greater than 70%, two from 50 to 70%, and two from 30 to 50%. This may indicate that the regions identified are not random and opens up prospects for their further in-depth analysis and determination of their significance in diagnosing schizophrenia. Testing carried out using data from another tomograph partially confirmed the greater significance of the identified regions for the classification of schizophrenia pathology, though ideal coincidence between the datasets from different tomographs was not achieved.</p>