Automatic Schizophrenia Identification from EEG Signals Using Dynamic Functional Connectivity and Deep Variational Autoencoders

Abstract

Schizophrenia is a severe neuropsychiatric disorder characterized by disruptions in perception, cognition, and behavior, making early and accurate diagnosis critical for effective treatment. Electroencephalography (EEG) has emerged as a promising non-invasive tool for capturing neural activity associated with schizophrenia. Recent advances in artificial intelligence (AI) have enabled automated analysis of complex EEG patterns, facilitating improved diagnostic accuracy. This study presents a comprehensive overview of AI techniques for automatic schizophrenia identification, focusing on dynamic functional connectivity (DFC) analysis and deep stack-augmented conditional variational autoencoders (DSA-CVAE). DFC captures temporal variations in brain connectivity, providing deeper insights into neural dynamics, while DSA-CVAE enhances feature representation through probabilistic latent modeling. The integration of these approaches allows for robust classification of schizophrenia by leveraging both spatial and temporal EEG characteristics. This paper reviews recent developments, compares methodologies, and highlights challenges such as data variability, model interpretability, and generalization across datasets. Furthermore, emerging trends including hybrid deep learning architectures, explainable AI, and multimodal integration are discussed. The findings indicate that combining advanced connectivity analysis with generative deep learning models holds significant potential for improving clinical decision support systems. However, further research is required to address scalability, reliability, and ethical concerns before widespread clinical adoption.