Detection of gait initiation Failure in Parkinson's disease based on wavelet transform and Support Vector Machine

Area of Science:

• Neuroscience
• Biomedical Engineering
• Clinical Neurology

Background:

• Gait Initiation Failure (GIF) is a subtype of Freezing of Gait (FOG) in Parkinson's disease (PD) patients, often causing falls.
• Understanding the neurobiological basis of GIF is challenging due to difficulties in clinical assessment.
• Electroencephalography (EEG) offers a potential method to study brain activity during GIF events.

Purpose of the Study:

• To develop and validate a novel EEG-based methodology for detecting GIF in PD patients.
• To improve the objective characterization of GIF phenomena using advanced signal processing techniques.

Main Methods:

• Utilized wavelet transform for feature extraction and Support Vector Machine for classification of GIF events.
• Applied Principal Component Analysis (PCA) to reduce EEG data dimensionality from 15 channels to 6 principal components (PCs).
• Employed Independent Component Analysis using Entropy Bound Minimization (ICA-EBM) on PCs for source separation to enhance GIF detection.

Main Results:

• The proposed methodology achieved 83.1% sensitivity, 89.5% specificity, and 86.3% accuracy in identifying GIF episodes.
• PCA successfully reduced data dimensions while retaining 93% of the information.
• ICA-EBM improved the detection of GIF events compared to normal gait initiation ('Good Starts').

Conclusions:

• The developed EEG analysis methodology shows promise as a non-invasive approach for improved GIF detection in Parkinson's disease.
• This technique could aid in better understanding and managing GIF in clinical settings.
• Further research is warranted to validate these findings in larger cohorts.