Fig. 2: Neural responses evoked by STN-DBS during different contact orientations.
a Source reconstruction of DBS-evoked cortical responses was conducted using unconstrained weighted minimum norm estimation and revealed robust medium- and long-latency sensorimotor (SM1) responses ipsilateral to the site of stimulation from 3–10 ms (shaded gray bar; top panel) and 16–26 ms (shaded gray bar; bottom panel) following DBS pulses denoted at time 0 ms, respectively. Time-domain average of the extracted SM1 response (i.e., dominant orientation time series) with time (in ms) on the x-axis and response amplitude (z-score) on the y-axis. All axes are fixed for each graph. b Grand averaged cortical patterns of medium- (top panel) and long-latency (bottom panel) responses in the ipsilateral SM1 across all patients, trials and contact orientations. c LME of medium- and long-latency SM1 response amplitude (i.e., <10 ms and ~20 ms, respectively) as a function of contact orientation (factor with 5 levels), controlling for acquisition order and total electrical energy delivered revealed stronger SM1 long-latency responses when clinically-effective contact directions were applied, while medium-latency SM1 responses were unchanged. Raincloud plots include a combined box plot (box edges: first 25th percentile quartile to third 75th percentile quartile; center line: median; data minima/maxima: whisker length), histogram distribution and individual scatter points of each evoked response. Significance is shown for post-hoc testing comparing optimal vs. all other non-optimal contact orientations at *pcorrected < 0.05.
