Non-equilibrium dynamics of the neocortex in Parkinson's disease

How dopamine depletion in Parkinson's disease (PD) and subsequent dopamine replacement therapy (DRT) affect brain activity is poorly understood. Typically, brain activity is analysed for its spectral properties and the temporal structure of the activity is often ignored. We quantified the time irreversibility of cortical activity in terms of entropy production rate (EPR). Using the Neural Estimator for Entropy Production (NEEP) algorithm on source-reconstructed resting-state magnetoencephalogram (MEG) data, we estimated the EPR in various brain regions of persons with PD (PwPD) (n = 17) and matched healthy controls (HC, n = 20). PwPD were recorded in two conditions: OFF and ON DRT (one hour after levodopa intake). HC were also recorded in two sessions separated by one hour. Motor symptoms were assessed with Movement Disorders Society-revised Unified Parkinson's Disease Rating Scale (MDS-UPDRS-III). Despite the lack of significant group differences in EPR between the HC and PD groups, we found a positive correlation between DRT induced improvement in motor symptoms (as measured by relative change in MDS-UPDRS-III scores (OFF-ON/OFF+ON)) and change in EPR. In particular, EPR changes in sensory (visual and auditory) regions and the information hub region called temporo-parieto-occipital junction were more strongly correlated with improvement in motor symptoms. Overall DRT tended to reduce EPR in PwPD and bring it close to the EPR values of HCs. Furthermore, PwPD with a higher EPR than HC showed better response to medication correlated with an increased EPR. Higher EPR in PwPD than HCs suggests that chronic dopamine loss alters local network interactions within cortical regions so as to reduce diversity of brain state transitions. Our analysis also opens a promising avenue to extract medication effects from non-invasively acquired cortical activity.