Frequency-Specific EEG burst dynamics in Parkinson's Disease and freezing of gait

Parkinson's disease (PD) disrupts motor networks, but conventional EEG spectral analyses may miss the transient, burst-like structure of pathological oscillations. Freezing of gait (FOG) is a disabling and poorly understood gait disturbance that may involve altered cognitive-motor control. This study aims to determine whether resting-state electroencephalography burst dynamics provide frequency-specific markers of PD and FOG.

Resting-state EEG was analyzed from 237 participants across three study sites, including 88 healthy controls and 149 individuals with PD assessed in the clinically defined ON-medication state. Participants with PD were also classified as having FOG (PDFOG +) or not having FOG (PDFOG -). Analyses were harmonized across sites using 11 common EEG channels. Burst dynamics were quantified in theta, alpha, low-beta, and high-beta bands using amplitude-envelope thresholding. Primary analyses focused on low- and high-beta burst dynamics at mid-frontal lead for healthy controls vs PD comparisons, and theta, low-beta, and high-beta burst dynamics for healthy control, PDFOG-, and PDFOG+ comparisons. FOG severity associations were examined using a study site-standardized FOGQ z-scores.

People with PD showed altered low-beta burst timing, characterized by more frequent, but shorter low-beta bursts compared with healthy controls. These effects were evident at the mid-frontal region and supported by exploratory topographic maps. In three-group analyses, low-beta burst rate and duration differed across groups, but adjusted PD-only models did not support robust PDFOG+ vs PDFOG- differences. In contrast, FOG severity was most consistently associated with theta amplitude-related burst metrics.

Resting-state EEG burst dynamics reveal dissociable oscillatory signatures in PD. Low-beta burst timing abnormalities reflect general PD-related motor network dysfunction, whereas theta burst amplitude may relate to FOG severity and cognitive-motor network burden.