Forged Channel: A Breakthrough Approach for Accurate Parkinson's Disease Classification using Leave-One-Subject-Out Cross-Validation

Abstract

This paper introduces a novel technique called "Forged Channel," which aims to comprehensively represent EEG signals in order to achieve accurate classification of Parkinson's disease. The forged channel method prepares EEG signals in a manner that allows a deep learning model to effectively perceive all EEG channels within a single input. By employing this approach alongside a convolutional neural network, an impressive accuracy of 90.32% was achieved using leave-one-subject-out cross-validation. This performance closely reflects real-world conditions, highlighting the superiority of our method compared to similar approaches.