Simultaneous recordings were obtained from the deep brain stimulation electrodes and 275-channel MEG (CTF/VSM MedTech, Vancouver, Canada) for all patients between two and six days postoperatively. At this time, continuous DBS for treatment purposes had not yet started. Data were online low-pass filtered at 600 Hz, sampled at 2400 Hz, and stored to disc. For a subset of cases (29–33) STN–LFPs were recorded using an EEG system separate from the MEG. MEG and LFPs were later temporally realigned and fused using a white noise reference signal recorded on both systems. A more comprehensive description of this methodology is provided in Oswal et al. (2015). Although the experiment also involved simple movement tasks for some of the patients, we here only analyse recordings obtained during periods of rest lasting about 3 min. Subjects were instructed not to move during the recordings and to focus their attention on a fixation cross without performing any explicit task. Separate experimental sessions were performed, for 13 out of 19 subjects on subsequent days, whilst the patient was either “ON” or “OFF” their normal dose of
