부산대학교 도서관

Parkinson's disease (PD) is a neurological disorder based on changes in dynamic brain activity, which can be partially ameliorated with invasive Deep Brain Stimulation. Galvanic vestibular stimulation (GVS), a non-invasive method, could potentially improve the motor symptoms of Parkinson's disease, but the mechanisms are unclear. Biomarkers based on the electroencephalogram (EEG) are being actively pursued. Here we examine the properties of EEG microstates as a potential GVS-sensitive EEG biomarker, whereby multichannel, broadband EEG signals are approximated by a sequence of discrete spatial patterns. We used the Microstate Analysis plugin for EEGLAB and compared the characteristics between healthy (n=20) and people with PD (n=22, stimulated/sham, and OFF Medication/ ON Medication). We extracted 25 Microstate related features from 4 different microstates (‘A’ - ‘D’) and examined their differences between groups (a healthy control group was considered as the reference to extract the feature values). Overall disease severity, as assessed by the clinical Unified Parkinson's Disease Rating Scale (UPDRS) Part 3, was predictable from microstate features. The duration of microstate A - selected by LASSO during UPDRS prediction- was significantly changed by both types of GVS stimuli (multi-sine 50–100 Hz (GVS1), and multi-sine 100–150 Hz (GVS2)), but not medication. The fraction of total recording time for microstate C, also a key feature in disease prediction, was found to be selectively affected GVS1 only. The above results suggest that GVS may provide benefits complementary to medication but in a stimulus-dependent manner. These results could potentially guide optimal GVS design in the pursuit of complementary therapies.