Diagnosis of Schizophrenia Patients in EEG Signals Using Multiple Optimizer-Convolutional Neural Network (CNN)

Schizophrenia is a mental disorder that impairs a person’s ability to think and behave clearly. The process of diagnosing schizophrenia can be time-consuming, depending on the specific protocols followed. One method employed in the diagnosis of schizophrenia involves the examination of Electroencephalogram (EEG) data. In this paper, a Schizophrenia detection by combined machine learning method and EEG was proposed. Initially, a dataset containing records of 81 patients obtained from Kaggle, later reduced to 30 patients through filtering. Data preprocessing involved utilizing a deep learning approach known as Convolutional Neural Network (CNN). Furthermore, six optimizers were employed to aid in the analysis, and the selection of the optimal model was based on the highest accuracy, sensitivity, specificity, precision, and F-measure values, as well as the loss and accuracy of the models. Based on the experimental results, the Root Mean Square Propagation optimizer yielded the highest values, with an accuracy of 0.8693, sensitivity of 0.66, precision of 0.74, specificity of 0.67, and an F-measure of 0.7.