Oscillatory Signatures of Parkinson's Disease: Central and Parietal EEG Alterations Across Multiple Frequency Bands

TL;DR

Parkinson's disease diagnosis remains largely clinical until substantial nigral degeneration; this study tests resting-state EEG as an early biomarker by applying artifact-cleaned EEG transformed into wavelet-based electrode-triplet images. A lightweight CNN classifies PD vs controls across multiple regions and frequency bands with leave-one-subject-out validation, highlighting central and parietal signatures, particularly in theta and alpha bands, and a distinctive gamma-band topography. The work supports the network-dysrhythmia view of PD and demonstrates that noninvasive EEG biomarkers, resistant to tremor contamination, could aid early detection and guide noninvasive neuromodulation strategies. Limitations include small sample size and cross-sectional design; future longitudinal studies in prodromal cohorts are needed.

Abstract

This study investigates EEG as a potential early biomarker by applying deep learning techniques to resting-state EEG recordings from 31 subjects (15 with PD and 16 healthy controls). EEG signals underwent preprocessing to remove tremor artifacts before classification with CNNs using wavelet-based electrode triplet images. Our analysis across different brain regions and frequency bands showed distinct spatial-spectral patterns of PD-related neural oscillations. We identified high classification accuracy (76%) using central electrodes (C3, Cz, C4) with full-spectrum 0.4-62.4 Hz analysis and 74% accuracy in right parietal regions (P8, CP6, P4) with 10-second windows. Bilateral centro-parietal regions showed strong performance (67%) in the theta band (4.0-7.79 Hz), while multiple areas demonstrated some sensitivity (65%) in the alpha band (7.8-15.59 Hz). We also observed a distinctive topographical pattern of gamma band (40-62.4 Hz) alterations specifically localized to central-parietal regions, which remained consistent across different temporal windows. In particular, we observed pronounced right-hemisphere involvement across several frequency bands. Unlike previous studies that achieved higher accuracies by potentially including tremor artifacts, our approach isolates genuine neurophysiological alterations in cortical activity. These findings suggest that specific EEG-based oscillatory patterns, especially in central and parietal regions and across multiple frequency bands, may provide diagnostic information for PD, potentially before the onset of motor symptoms.

Figures (7)

• Figure 1: Electrode placement according to the 10-20 international system showing the 32 electrodes used in this study. The light and dark shaded areas indicate electrode triplets with classification accuracy above 65% and above 70% correspondingly.
• Figure 2: Transformation of raw EEG signals RGB images. The signals are segmented into different window lengths (5, 10, and 20 seconds) followed by wavelet transformation to create 2D time-frequency representations, form which RGB images are formed by combining triplets of electrode channels.
• Figure 3: Architecture of the convolutional neural network used for classification. The model consists of three convolutional layers with batch normalization and ReLU activation, followed by max pooling, dropout for regularization, and a fully connected layer for binary classification (PD vs. healthy control). Different hyperparameters were optimized for each window length (5, 10, and 20 seconds) to maximize classification performance.
• Figure 4: Violin plot showing the distribution of classification accuracy for each electrode triplet using 20-second windows and the full frequency spectrum.
• Figure 5: Violin plots showing the distribution of classification accuracy using 10-second windows for each electrode triplet across different frequency bands: delta (top-left), theta (top-right), alpha (mid-left), beta1 (mid-right), beta2 (bottom-left), gamma (bottom-right).