EEG Epileptic Data Classification Using the Schrodinger Operator's Spectrum.

Area of Science:

• * Neuroscience
• * Biomedical Engineering
• * Quantum Computing Applications

Background:

• * Epilepsy affects over 65 million people globally, characterized by recurrent seizures.
• * Manual Electroencephalogram (EEG) analysis for epilepsy diagnosis is time-consuming and requires specialized expertise.
• * There is a growing need for automated systems to improve diagnostic efficiency and accuracy.

Purpose of the Study:

• * To develop an autonomous computer-aided system for epileptic seizure detection using EEG data.
• * To explore the efficacy of Semi-Classical Signal Analysis (SCSA) and nonlinear dynamical features for EEG classification.
• * To optimize feature selection and employ machine learning classifiers for high-accuracy diagnosis.

Main Methods:

• * Extraction of features using the Semi-Classical Signal Analysis (SCSA) method, a quantum-inspired signal processing technique.
• * Integration of nonlinear dynamical features known for characterizing neural activity.
• * Application of hyperparameter optimization, correlation analysis, and feature selection.
• * Classification using five distinct machine learning algorithms on the Bonn University EEG database.

Main Results:

• * The proposed approach achieved classification accuracy of 93% and above across all tested machine learning classifiers.
• * Demonstrated the effectiveness of SCSA and nonlinear dynamical features in characterizing epileptic EEG signals.
• * Validated the system's performance on a well-established public EEG dataset.

Conclusions:

• * The developed system offers a promising, highly accurate, and automated solution for epileptic seizure diagnosis.
• * Contributes to reducing diagnosis time and potential errors in clinical practice.
• * Highlights the potential of quantum-inspired signal processing in neurological disorder diagnostics.