Harnessing EEG Signals to Detect Schizophrenia: A Deep Learning Approach

This research paper stages a deep convolutional neural network-based approach for the detection of a complex and debilitating mental disorder, schizophrenia. This mental disorder is particularized through a spectrum of pointers including delusions, hallucinations, perplexing thoughts, and impaired speech. The study focuses on utilizing a Convolutional Neural Network (CNN) architecture explicitly designed for EEG data analysis. EEG signals from 14 healthy subjects and 14 schizophrenic patients were used to train and validate the CNN model. The proposed CNN architecture incorporates multiple Conv1D layers with varying filter sizes and activation functions, allowing it to effectively capture intricate temporal patterns within the EEG signals. The architecture is augmented with pooling layers, enabling hierarchical feature extraction and dimensionality reduction. Elements like LeakyReLU activations and Average Pooling1D layers are integrated to achieve optimal feature extraction. The CNN is designed to output probabilistic classifications for each input EEG signal, enabling a binary diagnosis prediction. The efficacy of the CNN model is validated on the standard assessment metrics of accuracy, specificity, and sensitivity. The accuracy generated by the proposed CNN model is 97.55%. Thus, this research highlights the significance of combining an advanced machine learning approach with EEG data analysis, ultimately contributing to the advancement of early diagnosis and intervention strategies in Schizophrenia.