Novel automated PD detection system using aspirin pattern with EEG signals.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Biology

Background:

• Parkinson's disease (PD) significantly impacts global quality of life.
• Electroencephalogram (EEG) signals combined with machine learning offer automated PD detection methods.

Purpose of the Study:

• To propose a novel graph-based aspirin model for accurate PD detection using EEG signals.
• To investigate the feature generation capabilities of chemical graphs in the context of PD diagnosis.

Main Methods:

• A multi-level feature generation phase incorporating a new aspirin pattern, statistical moments, and maximum absolute pooling (MAP).
• Selection of discriminative features using Neighborhood Component Analysis (NCA).
• Classification of PD using k-Nearest Neighbors (kNN) and iterative majority voting.

Main Results:

• The proposed model achieved high classification accuracies of 93.57% for Case 1 and 95.48% for Case 2.
• Leave-One-Subject-Out (LOSO) validation was employed for robust result calculation.
• A public dataset was utilized for model development and validation.

Conclusions:

• The developed automated PD detection model demonstrates high accuracy.
• The model is suitable for further testing with diverse EEG datasets.