Predicting the Onset of Freezing of Gait Episodes in Patients with Parkinson's Disease Using EEG

Parkinson's disease is a neurodegenerative disorder that affects movement control, causing symptoms such as tremors, stiffness, and slowness. However, as the disease progresses, most patients also experience freezing of gait (FOG), a sudden, debilitating episode characterized by an inability to walk, which increases the risk of falls and injuries. Currently, no effective treatments exist for FOG, so detecting the onset of a freezing episode with sufficient lead time could enable preventative measures to avoid falls and injuries. This study proposes a method to detect the onset of FOG in PD patients by analyzing brainwave patterns from ambulatory electroen-cephalogram (EEG) signals. Our approach utilizes spectral analysis to capture frequency-specific patterns associated with both onset of FOG episodes and non-FOG. These features serve as input to a machine learning algorithm to discern subtle changes indicative of impending FOG. A dataset of EEG recordings from PD patients experiencing FOG episodes is utilized to train and validate the proposed approach. Our method achieves an AUC of 0.825, demonstrating the potential to detect the onset of FOG episodes which could pave the way for timely interventions to alleviate or prevent FOG episodes. The proposed method holds promise for enhancing the overall management and care of individuals with PD, ultimately contributing to improved patient outcomes and a better quality of life.