We first focus our account on the analyses conducted for the LFP recordings obtained from the DBS electrodes, those for the MEG data are described later. DBS was switched off during the experiments, allowing for the acquisition of LFPs from all four contact points of the STN electrodes. These were off-line converted to a bi-polar montage between adjacent contacts, resulting in 3 channels per site. This was done to limit volume conduction of signals form distal sources. 50 Hz line noise and its harmonics up to 600 Hz were filtered out using 4th order Butterworth bandstop filters with cutoff frequencies ±0.5 Hz around the centre frequencies. We subsequently divided the continuous time series into epochs of 3.41 s and discarded epochs where the peak-to-peak LFP amplitude exceeded 100 μV. On average, 60 epochs (range 13–106) remained per STN site for further analyses. We included all available data in our analyses. This amounted to a dataset comprising recordings from 53 hemispheres with patients OFF medication and 48 hemispheres with patients ON medication.
