# EEG dynamical features during variable-intensity cycling exercise in Parkinson’s disease

## TL;DR

This study explores how brain activity changes with exercise intensity in Parkinson’s disease patients using EEG, identifying neural markers that could help guide effective exercise regimens.

## Contribution

The study introduces EEG-derived dynamical features as potential biomarkers to assess and guide exercise intensity in Parkinson’s disease.

## Key findings

- A significant MCCA component showed a monotonic relationship between EEG features and exercise intensity.
- Nine EEG features demonstrated significant trends with intensity, with six increasing and three decreasing.
- PD patients showed more rigid neural responses compared to healthy controls, who exhibited greater variability.

## Abstract

Exercise is increasingly recognized as a beneficial intervention for Parkinson’s disease (PD), yet the optimal type and intensity of exercise remain unclear. This study investigated the relationship between exercise intensity and neural responses in PD patients, using electroencephalography (EEG) to explore potential neural markers that could be ultimately used to guide exercise intensity.

EEG data were collected from 14 PD patients (5 females) and 8 healthy controls (HC) performing stationary pedaling exercises at 60 RPM with resistance adjusted to target heart rates of 30, 40, 50, 60, and 70% of maximum heart rate. Subjects pedaled for 3 min at each intensity level in a counterbalanced order. Canonical Time-series Characteristics (Catch-22) features and Multi-set Canonical Correlation Analysis (MCCA) were utilized to identify common profiles of EEG features at increasing exercise intensity across subjects.

We identified a statistically significant MCCA component demonstrating a monotonic relationship with pedaling intensity. We have discovered nine features which show significant trends across intensity (p-value<0.01), and the dominant feature in this component was Periodicity Wang (p-value<0.0001), related to the autocorrelation of the EEG. Analysis revealed a consistent trend across features: six features increased with intensity, indicating heightened rhythmic engagement and sustained neural activation, while three features decreased, suggesting reduced variability and enhanced predictability in neural responses. Notably, PD patients exhibited more rigid, consistent response patterns compared to healthy controls (HC), who showed greater flexibility and variability in their neural adaptation across intensities.

This study highlights the feasibility of using EEG-derived features to track exercise intensity in PD patients, identifying specific neural markers correlating with varying intensity levels. PD subjects demonstrate less inter-subject variability in motor responses to increasing intensity. Our results suggest that EEG biomarkers can be used to assess differing brain involvement with the same exercise of increasing intensity, potentially useful for guiding targeted therapeutic strategies and maximizing the neurological benefits of exercise in PD.