Gyanendra Kumar Verma

Schizophrenia is a complex neuropsychiatric disorder that significantly impacts cognitive functioning and quality of life. In this study, we propose a deep learning-based diagnostic framework leveraging electroencephalogram signals for automated SZ detection. The raw EEG data were pre-processed and transformed into time–frequency representations using the Continuous Wavelet Transform with the Morlet wavelet to generate informative scalograms. These scalograms were resized and utilised as input to a two-dimensional Convolutional Neural Network, designed to extract deep discriminative features for binary classification. In channel-wise analysis, where each EEG channel was analysed independently, yielding a classification accuracy of 95.91%, and subject-wise analysis, where aggregated channel information was used for subject-level classification, achieving 88.24% accuracy. The findings indicate that the integration of CWT-based feature extraction with deep learning architectures provides a robust and interpretable approach for EEG-based Schizophrenia detection.