Automatic identification of schizophrenia based on EEG signals using dynamic functional connectivity analysis and 3D convolutional neural network

• Published in: Computers in biology and medicine Vol. 160; p. 107022
• Main Authors: Shen, Mingkan, Wen, Peng, Song, Bo, Li, Yan
• Format: Journal Article
• Language: English
• Published: United States Elsevier Ltd 01.06.2023; Elsevier Limited
• Subjects: 3D convolutional neural network; Accuracy; Alcoholism; Algorithms; Artificial neural networks; Brain - diagnostic imaging; Brain mapping; Brain research; Classification; Connectivity analysis; Cross mutual information; Datasets; Deep learning; Default mode network; Discriminant analysis; EEG; Electroencephalography; Electroencephalography - methods; Emotions; Functional magnetic resonance imaging; Humans; Identification; Internal Medicine; Machine learning; Medical research; Mental disorders; Neural networks; Neural Networks, Computer; Other; Performance evaluation; Schizophrenia; Schizophrenia - diagnostic imaging; ScZ; Signal processing; Support vector machines; Temporal lobe; Wavelet transforms; Default mode network; ScZ; Cross mutual information; EEG; 3D convolutional neural network
• ISSN: 0010-4825; 1879-0534; 1879-0534
• DOI: 10.1016/j.compbiomed.2023.107022

Schizophrenia (ScZ) is a devastating mental disorder of the human brain that causes a serious impact of emotional inclinations, quality of personal and social life and healthcare systems. In recent years, deep learning methods with connectivity analysis only very recently focused into fMRI data. To explore this kind of research into electroencephalogram (EEG) signal, this paper investigates the identification of ScZ EEG signals using dynamic functional connectivity analysis and deep learning methods. A time-frequency domain functional connectivity analysis through cross mutual information algorithm is proposed to extract the features in alpha band (8–12 Hz) of each subject. A 3D convolutional neural network technique was applied to classify the ScZ subjects and health control (HC) subjects. The LMSU public ScZ EEG dataset is employed to evaluate the proposed method, and a 97.74 ± 1.15% accuracy, 96.91 ± 2.76% sensitivity and 98.53 ± 1.97% specificity results were achieved in this study. In addition, we also found not only the default mode network region but also the connectivity between temporal lobe and posterior temporal lobe in both right and left side have significant difference between the ScZ and HC subjects. •The time-frequency domain functional connectivity calculated by continuous wavelet transform (CWT) and cross mutual information (CMI) is firstly used in schizophrenia identification and the frequency resolution is selected in 1 Hz in this experiment.•Sliding window technique is proposed to extend the functional connectivity to time-varying functional connectivity for exploring dynamic properties of resting-state function connectivity in EEG.•To reduce the computational cost, graph theory measures of complex brain network analysis are used to select brain rhythms and find alpha band (8–12 Hz) is the significance frequency band for schizophrenia (ScZ) identification work.•The 3D-CNN models are applied to classify the ScZ subjects and health control subjects and achieved a result of 97.74 ± 1.15% in accuracy, 96.91 ± 2.76% sensitivity and 98.53 ± 1.97% specificity.•Furthermore, we analysed the CMI values in the whole connectivity and found not only the default mode network (DMN) region but also the connectivity between temporal lobe and posterior temporal lobe in both right and left side has significant difference between the ScZ and healthy control (HC) subjects.