SynaptiQ: A Hybrid Neuro-Analytical Framework for Multidimensional Psychiatric Disorder Prediction from EEG Signals

Abstract

Neurological and psychiatric disorders are increasingly impacting global health, making early and accurate diagnosis essential for effective treatment and improved patient outcomes. Electroencephalography (EEG) signals provide a non-invasive and information-rich source for analyzing brain activity; however, their high dimensionality and complex patterns make automated classification a challenging task. Many existing EEG-based diagnostic approaches face limitations such as poor generalization, lack of subject independence, and difficulty in capturing complex nonlinear relationships within brain signals, resulting in reduced accuracy and reliability. To address these challenges, this work proposes a multiclass machine learning (ML) framework for the simultaneous classification of epilepsy, migraine, and schizophrenia using EEG data. The system incorporates systematic preprocessing techniques including data cleaning, missing value imputation, normalization, and label encoding, followed by stratified train–test splitting to ensure balanced and unbiased learning. Multiple models, including Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), and Decision Tree (DT), are implemented and evaluated alongside a Hybrid Graph Neural Network (HGNN) model, where the HGNN component is realized using a Tree-based Generalized Additive Model (TreeGAM) to effectively capture complex nonlinear feature interactions. The system is deployed through a Flask-based web application with secure authentication and role-based access control, allowing administrators to manage training and analysis while users perform predictions. Performance evaluation using metrics such as accuracy, precision, recall, and F1-score demonstrates the effectiveness of the proposed framework, with the hybrid model showing improved predictive capability. The system provides a scalable, reliable, and practical decision-support tool for data-driven mental health diagnosis