Neurodynamic Characterization and Prediction of Schizophrenia Using Echo State Networks with Serotonin Modulation: A Temporal and Frequency Band Analysis Approach

Abstract Schizophrenia is characterized by significant cognitive dysfunctions, with serotonin playing a crucial role in modulating neural processes. Analyzing the impact of serotonin on EEG patterns can provide important insights into distinguishing schizophrenic patients from healthy individuals. This study integrates serotonin-inspired modulation into Echo State Networks (ESNs) to model the nonlinear dynamics of EEG data from schizophrenic patients, with a focus on key brain regions such as the frontal lobe and medial prefrontal cortex (mPFC). EEG data were preprocessed (0.1–40 Hz), ICA filtered, segmented, and analyzed using ESNs, with serotonin modulation targeting 5 HT1A and 5-HT2A receptors to simulate effects on frontal and mPFC regions. Principal Component Analysis (PCA) and K-means clustering were used to classify schizophrenic and healthy samples. Results showed that EEG patterns in schizophrenia exhibited elevated delta and gamma band activity, with diminished alpha and beta band activity. Serotonin modulation enhanced the performance of ESNs by reducing noise and improving predictive accuracy, particularly in the frontal regions. The incorporation of serotonin modulation into ESNs offers a deeper understanding of the nonlinear dynamics of brain activity in schizophrenia and provides valuable insights into disease-specific EEG features, potentially advancing diagnostic approaches.