Fig. 5: Deep brain stimulation alleviates burst related reductions of acceleration during free drawing.

A Top: For visualization, velocity signals within windows of 250 ms before and after the occurrence of a burst were normalized by the instantaneous velocity during burst onset and averaged within each subject. Average velocity signals and the standard error of the mean (shaded area) for all drawing and stimulation conditions (blue = free drawing without stimulation; red = free drawing with stimulation; green = template-guided drawing without stimulation; orange = template-guided drawing with stimulation) are shown. We calculated the average acceleration within the pre and post intervals for each burst. Two-sided linear mixed-effects models showed main effects of time interval (P < 0.001), and stimulation condition (P = 0.001) and interaction effects for stimulation condition * drawing condition (P = 0.007), and drawing condition *time interval (P = 0.019). Two-sided post hoc tests demonstrated that only bursts occurring during free drawing are accompanied by a reduction in acceleration (as can be seen as a flattening of the blue velocity signal after the occurrence of a burst, P = 0.002). When applying DBS, the acceleration around beta bursts was generally increased during free drawing, suggesting that stimulation modifies their immediate impact on velocity dynamics (P = 0.008, two-sided post hoc). When drawing with a template, acceleration was not affected by the occurrence of a burst (P = 1, two-sided post hoc). Bottom: modeled beta amplitude dynamics demonstrating the definition of the burst onset as the time point when the amplitude crosses the 75th percentile threshold. B Average velocity dynamics and the standard error of the mean (shaded area) for the two control conditions showing that the velocity signal does not change its slope after the occurrence of linearly or randomly defined onsets (top: burst onsets defined in 50 ms steps across every trial; bottom: randomly assigned timestamps). Two-sided linear mixed-effects models did not show any significant effects of time interval (P = 0.143 and P = 0.256, linear and random assignment, respectively), stimulation condition (P = 0.456 and P = 0.193) or interaction between stimulation condition * time interval (P = 0.913 and P = 0.867), highlighting that the average acceleration after the defined onsets is not significantly different from before. Source data are provided as a Source Data file.