Figure 3.

Effect of bursts before and overlapping with the GO cue on the amplitude of the peak velocity and impact of burst detection threshold. A, Mean peak velocity in burst trials normalized (z score) to the mean velocity of all trials for all subjects. A negative value indicates a reduction of peak velocity in burst trials. Trials are divided according to the presence of a burst in a 600 ms window before the GO cue where bursts are only included if more than half of their duration falls in the time window. Bursts were defined with the default threshold of 75th percentile. B, Impact of burst detection threshold on the peak velocity reduction. For each subject, the velocity peak of each trial is normalized (z scores) as described for A. C, Estimated effects and 95% CIs derived from the linear mixed-effects models testing the impact of bursts occurring before or overlapping with the GO cue on peak velocity. Burst detection thresholds stop at 85th as too few trials with bursts were identified for the next 90th threshold. For the modeling, the peak velocities were power transformed (see Materials and Methods). *Significant model (p < 0.05).