Fig. 1: Methods.

a Patient timeline consisting of at least one out of two LFP recordings. b IPG bipolar recording montage. LFPs were derived from DBS leads with eight contacts: ring contacts at levels 0 and 3 and segmented contacts at levels 1 and 2 (1 A/B/C, 2 A/B/C). IPG bipolar recording montage of 15 bipolar channels recorded in three channel groups: ring, segmental-2L and segmental-1L LFP. c LFP processing. Externalized LFPs (blue) were re-referenced in two formats: EXT-1 with neighboring ring channels as reference for segmented channels, reflecting the most accurate “real” beta distribution from our externalized recordings; EXT-2 was re-referenced to simulate the IPG bipolar recording montage. To estimate the beta distribution from IPG-bipolar recordings (red) and pseudo-IPG-bipolar EXT-2 recordings (blue), three methods were applied. d Short-term stability of maximal beta power was analyzed by (1) selecting the maximal beta channel of 2-min EXT-2 recordings (#1, separately for ring and segmental LFPs) and (2) extracting multiple 20-s windows with 50% overlap. Beta ranks (ranks #1 to #3 for ring channels and ranks #1 to #9 for segmental channels) were extracted from all short windows. e The Euclidean method utilizes all available n = 9 segmental LFPs weighted by the distance of their central position between the recording contacts and the contact of interest, for instance, 1B. f The Strelow et al. method utilizes n = 2 segmental-1L channels and weights them by the distance of the contact partners. g The Busch et al. method combines one ring LFP with the segmental-2L LFPs. Pseudo-monopolar power spectra from an example hemisphere are depicted on the right for each method.