Pathological oscillations are a hallmark of neural activity in Parkinsons disease (PD). Time-averaged analyses are usually employed to study changes in spectral connectivity with and without dopaminergic intervention in PD. This prevents differentiating the pathological vs physiological nature of dynamically evolving oscillatory activity serving multiple functional roles. Using a Hidden Markov Model on combined STN-LFP and whole-brain MEG data from 17 PD patients we discovered three distinct network activity patterns. One network was related to adverse effects of increased dopamine, a second one maintained ON-medication spatio-spectrally selective cortico-STN connectivity and finally, a local STN-STN network emerged which indicated the inability of L-DOPA to modify local basal ganglia activity. Temporally we found that, ON medication, the cortico-STN and the STN-STN network increased in duration whereas the cortico-cortical network occurred less frequently. Our results provide a spectrally diverse and spatially specific understanding of transient network connectivity in PD on a whole-brain level, disambiguating temporal and spatial changes of the underlying networks. By providing electrophysiological evidence for the differential effects of L-DOPA intervention in PD, our findings open further avenues for electrical and pharmacological intervention in PD.
bioRxiv Subject Collection: Neuroscience