Task-specific dynamical entropy variations in EEG as a biomarker for Parkinson's disease progression

Area of Science:

• Neuroscience
• Cognitive Science
• Biomedical Engineering

Background:

• Parkinson's disease (PD) disrupts neuronal mechanisms essential for dynamic brain adaptation.
• This leads to impairments in motor coordination and cognitive functions in PD patients.
• Understanding these disruptions is key to tracking PD progression and developing interventions.

Purpose of the Study:

• Investigate neuronal dynamics during a lower-limb pedaling task in healthy controls (HC), PD patients, and PD patients with freezing of gait (PDFOG).
• Analyze changes in dynamical entropy of EEG signals to characterize task-specific neural adaptations.
• Explore the potential of entropy-derived features for classifying PD subtypes.

Main Methods:

• Utilized electroencephalography (EEG) to record brain activity during a pedaling task.
• Analyzed permutation entropy and Vasicek's entropy variability in EEG signals.
• Employed Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM) classifiers for group discrimination.

Main Results:

• PD and PDFOG patients showed decreased permutation entropy in frontal and parietal regions, indicating altered information processing.
• Reduced Vasicek's entropy variability was observed in occipital and left frontal regions for both PD groups, suggesting diminished cognitive resource allocation.
• Entropy-derived features achieved up to 96.15% accuracy in distinguishing HC from PDFOG patients.

Conclusions:

• The dynamical entropic framework offers a novel method to assess neural complexity changes during task performance.
• This approach can reveal subtle cognitive-motor impairments in Parkinson's disease.
• Entropy analysis holds potential for tracking PD diagnosis and progression, complementing resting-state analyses.