Fig. 2: Results.

a Validation of pseudo-monopolar power estimation methods. Estimated pseudo-monopolar beta distributions (left) and maximal beta contacts (#1, #2, right) using three methods (Euclidean, Busch et al., Strelow et al.) applied to EXT-2 recordings were compared to the “real” EXT-1 outcomes of the same hemispheres. Left: Most hemispheres showed positive correlation coefficients for all three methods. Right: Euclidean and Strelow et al. methods identified maximal beta contacts significantly more often than chance. b Agreement between methods. Estimated beta distributions using the Euclidean and Strelow et al. methods showed higher inter-method correlation than comparisons involving the Busch et al. method. c Agreement of maximal beta contacts with clinically active stimulation contacts. Maximal beta contacts derived from the “real” EXT-1 distribution (blue bars) or from estimated distributions based on IPG recordings (red bars) were chronically active for stimulation at 12 months post-surgery in ~37–67% of hemispheres. d Short-term stability of beta ranks. Two-minute maximal beta channels were analyzed for ring (top) and segmental channels (bottom) from n = 30 leads (one additional hemisphere was excluded due to its recording being slightly shorter than 2 min). Left: Boxplots show that the coefficient of variance (CV) of beta ranks from 20-s windows was lower than from 5-s windows, and comparable to 30-s windows. In boxplots: black cross = group mean; horizontal line = median; vertical black line = data range; box = interquartile range; diamonds = outliers. Right: Proportion of #1 channels correctly identified from randomly selected 20-s windows across 1000 iterations. Each iteration involves a certain number of randomly selected windows (mean ± CI). An example is provided showing the proportions of correctly identified #1 channels from two selected windows. For ring channels, ≥3 windows ensured about 90% certainty (red line) in correctly identifying the #1 channel. For directional channels, ≥3 windows ensured 60% certainty.