Longitudinal changes of resting-state networks in Parkinson's disease

Abstract

Deep brain stimulation (DBS), but also the sole implantation of the electrodes and dopaminergic medication, can reduce symptoms in Parkinson's disease (PD) patients. Furthermore, an effect on network activity is known for all three options separately. However, long-term effects have rarely been investigated. Therefore, in the present study, we focus on the long-term impact of dopaminergic medication on whole-brain network activity following DBS electrode implantation. Therefore, we extracted resting state networks (RSNs) of 20 PD patients (4 females, (59.00 ± 9.72 years) from magnetoencephalography data. We recorded 30 min of resting-state activity two days before and one year after implantation of the electrodes with and without dopaminergic medication, but DBS was turned off. RSNs were obtained based on the phase-amplitude coupling between a low-frequency phase and a high gamma amplitude and examined for differences between conditions (i.e., pre- vs. post-surgery). We identified three RSNs across all conditions: sensory-motor, visual, and frontal. Each RSN was selectively altered due to a year of disease progression - while patients being treated with dopaminergic medication and DBS. In line with previous literature, we focus on longitudinal changes in RSNs over time after electrode implantation, acknowledging that chronic DBS treatment and other factors may confound the interpretation of these changes. In addition, the alterations found were RSN specific, as dopaminergic medication showed a greater impact on the frontal RSN, and the longitudinal factor expressed by the disease progression was more severe in alterations in the SMN and the visual RSN.

Keywords: MEG; PAC; PD; Resting-state network; Stun effect.

Fig. 1

Resting state networks investigated in four different Parkinson’s disease conditions The resting state networks (RSN) under consideration (sensory-motor, visual, and frontal) are arranged in blocks one below the other. The rows of each block display the network for the recordings before electrode implantation (first) OFF and ON medication, and the recordings after (second) electrode implantation ON and OFF medication are below. The color scale marks the coupling strength from 0 to 1, with a warmer color indicating a higher coupling strength. No threshold was applied. It can be seen that the focus of the coupling strength for the individual RSNs is located in different regions of the brain. When comparing the conditions, it is also noticeable that the topography of the RSN changes.

Fig. 2

Alteration in the resting state networks The alterations in the resting state networks (RSN) under consideration (sensory-motor, visual, and frontal) are displayed in blocks one below the other. The RSNs were compared after recordings from Parkinson’s patients after electrode implantation (second) and before (first) and with medication (ON) and without (OFF). Areas belonging to either one and/or both recording conditions are marked in white. Red indicates a significantly higher coupling strength for the former measurement group and blue for the latter one. Significance is given at a p-value below 0.05 after false discovery rate correction for the number of vertices, networks, and conditions based on a two-sided Student’s t-test (HC vs. OFF/ON independent t-test, ON vs. OFF paired t-test). For the sensory-motor RSN in particular, alterations were only present in the comparisons second vs. first recording but not OFF vs. ON. Also, for the visual RSN, alterations were only present in the comparisons second vs. first recording but not OFF vs. ON but more severe than for the sensory-motor RSN. However, for the frontal RSN, alterations were more prominent for the OFF vs. ON comparisons. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Spatial correspondence between resting-state networks of healthy controls and Parkinson's patients The spatial correspondence between resting-state networks calculated with the phi-coefficient based on the thresholded and then binarized coupling strength between the RSNs of PD conditions (second-OFF, line indicator: asterisk; second-ON, x; first-OFF, +; first-ON, square) and the healthy controls is plotted for each jackknife run (x-axis). The number of patients and, therefore, jackknife runs differed between conditions.