fNIRS-Based Graph Frequency Analysis Quantified the Effect of Dopaminergic Medication on Gait in Parkinson's Disease

Dopaminergic treatment effectively improves gait function in patients with Parkinson's disease (PD). Previous studies have mostly used clinical scores to evaluate the efficacy of dopaminergic medication on gait in patients with PD. However, clinical scores are not able to assess changes in brain function. A method of graph frequency analysis based on functional near-infrared spectroscopy (fNIRS) was designed to quantify the effect of dopaminergic medication on gait in patients with PD. Twelve participants completed a gait task using fNIRS in both medication OFF and ON states. A functional brain network was constructed by calculating Pearson's correlation coefficient for each pair of fNIRS channels. The network was represented as a graph and analyzed in the graph frequency domain using the graph Laplacian eigenvectors, resulting in the eigenvector matrix. To measure functional interaction between and within brain regions in specific graph frequency ranges, total variations (TV) were determined and applied. The experimental results indicate that in the ON state, the frontal lobe exhibited an increase in low-frequency components and a reduction in high-frequency components. In contrast, the parietal lobe displayed the opposite pattern. A correlation was found between the change in frontal and parietal functional interaction and the change in gait. This study shows that the graph frequency analysis method based on fNIRS can effectively quantify the impact of dopaminergic treatment on gait in PD.