Classification of EEG signals to detect Parkinsons Disease using a computationally method

The use of EEG signals in neuroscience research has grown significantly due to their ability to capture and measure various brain phenomena. However, the analyses of these signals are in somehow challenge due to their random nature and susceptibility to noise, often requiring extensive computational resources. To tackle these challenges, a new approach presented, called Forged Channels combined with Smoothed Pseudo Wigner-Ville Distribution (SPWVD) transformation. This new method enables accurate analysis of EEG signals while maintaining computational cost efficient. By utilizing this approach alongside a CNN classifier, we achieved promising results, with a 90% accuracy in classifying Parkinsons Disease through Leave-One-Subject-Out Cross-Validation which is an strong evidence for the effectiveness of our proposed methodology in representing EEG signals.