SzHNN: A Novel and Scalable Deep Convolution Hybrid Neural Network Framework for Schizophrenia Detection Using Multichannel EEG

In the field of neuroscience, brain activity measurement and analysis are considered crucial areas. Schizophrenia (Sz) is a brain disorder that severely affects the thinking, behavior, and feelings of people worldwide. Thus, an accurate and rapid detection method is needed for proper care and quality treatment of the patients. Electroencephalography (EEG) is proved to be an efficient biomarker in Sz detection as it records brain activities. This article aims to improve the performance of EEG-based Sz detection using a deep-learning approach in remote applications. A hybrid deep-learning model identified as schizophrenia hybrid neural network (SzHNN), which is a combination of convolutional neural networks (CNNs) and long short-term memory (LSTM), has been proposed wherein the CNN for local feature extraction and LSTM for classification is utilized. In this article, the proposed model has been compared with several deep-learning and machine-learning-based models. All the models have been evaluated on two different datasets wherein dataset 1 consists of 19 subjects and dataset 2 consists of 16 subjects. The proposed model is also implemented with the Internet-of-Medical-Things (IoMT) framework for smart healthcare and remote-based applications. Several experiments have been conducted using various parametric settings on different frequency bands and different sets of electrodes on the scalp. Based on all the experiments, it is evident that the proposed hybrid model (SzHNN) provides the highest classification accuracy of 99.9% compared to other implemented models and existing models of previous papers. The proposed model overcomes the influence of different frequency bands and shows a better accuracy of 96.10% (dataset 1) and 91.00% (dataset 2) with only five electrodes. Subject-wise testing is also done for SzHNN, which proposes an accuracy of 90.11% and 89.60% for datasets 1 and 2, respectively.