Identify the brain abnormalities from resting-state fMRI data for classification of schizophrenia

Due to the lack of effective biological markers, the development of diagnosis and treatment methods for schizophrenia and other related brain diseases has been slow in the past decades. Based on resting-state functional magnetic resonance imaging, this study simultaneously searched for brain abnormalities in patients with schizophrenia from two aspects of spontaneous neural activity in local brain regions and brain network topology. Based on this, an automatic computer diagnosis algorithm was designed to classify patients with schizophrenia and normal controls. The results showed that the spontaneous neural activity of patients in thalamus, postcentral gyrus and other regions was significantly reduced; The results showed that the global efficiency of brain network was significantly increased, the small world was slightly reduced but not significantly, and the local topological properties of parahippocampal gyrus, temporal pole: middle temporal gyrus were significantly changed. The features extracted from abnormal brain regions could effectively distinguish patients from normal controls, and the highest classification accuracy was 80.17%. The findings of this study can help to identify effective biomarkers for clinical diagnosis and provide insights into the neurophysiological mechanisms of schizophrenia. The proposed classification algorithm can also be used to assist the diagnosis of schizophrenia.