Results

Reaction Time

Premotor RT as a function of TMS application time (−500, −300, −150, −100, −50, and 0 ms) for each RT task (simple, choice, and go/no‐go) is presented in Figure 2. Analysis revealed a significant main effect of RT task, F(2,28.7) = 55.014, p < 0.001, indicating that RT for the simple RT task (204 ms, SE = 6.0 ms) was significantly shorter than both go/no‐go task (260 ms, SE = 5.3 ms) and the choice RT task (251 ms, SE = 6.3 ms), whereas there was no difference between the latter two. There was also a main effect of TMS time, F(5,3059.7) = 2.544, p = 0.026, with post hoc tests revealing that RT was significantly shorter (10 ms, SE = 3.4) when TMS was applied at 150 ms prior to the go versus at the go‐signal (p = 0.036). Finally, there was no significant interaction between the factors (p = 0.864).

FIGURE 2: Premotor reaction time (RT) as a function of TMS stimulation time for each RT task (simple RT, choice RT, and go/no‐go RT). Fixed effects means and 95% CI are shown with large black markers and error bars. Individual subject estimated mean RTs are shown as small red markers, and violin plots represent the distribution of raw data. Note: RT was significantly shorter in the simple RT condition compared to choice and go/no‐go. There was also a significant reduction in RT in each task when TMS was applied at the −150‐ms time point as compared to time zero (shown with asterisk *).

MEP Amplitude

Integrated MEP amplitude when TMS was applied at rest (prior to beginning any RT task; n = 12) was compared to MEP amplitude when TMS was applied at the warning signal in each of the RT tasks (n = 30). This was done to determine if corticospinal excitability was differentially elevated from baseline between the RT tasks (Figure 3). Analysis showed a significant effect of condition, χ 2(3) = 146.22, p < 0.001, with post hoc comparisons indicating that while MEP amplitude was significantly larger in all three RT tasks compared to baseline (all p‐values < 0.001), there was no difference in amplitude between the RT tasks (all p‐values > 0.743).

FIGURE 3: Motor‐evoked potential (MEP) integrated area at baseline (resting; gray) and at the warning signal (−500 ms) for each reaction time (RT) task (green = simple RT; blue = choice RT; red = go/no‐go RT). Fixed effect means and 95% CI are shown as large markers and error bars; individual subject estimated mean MEP amplitudes are shown as small gray markers, and violin plots represent the distribution of raw data. Means are back‐transformed from the log scale. Note that MEP amplitude was significantly (*) larger for all three RT tasks compared to baseline.

Integrated MEP amplitude as a function of TMS time during each RT task is presented in Figure 4. Significant main effects were found for TMS time, χ 2(5) = 267.587, p < 0.001, and RT task, χ 2(2) = 16.980, p < 0.001, as well as a significant interaction between TMS time and task, χ 2(10) = 18.664, p = 0.045. Post hoc tests indicated that in each RT task, MEP amplitude was significantly smaller (all p‐values < 0.01) when TMS was applied at −150, −100, and −50 ms, and at the go‐signal, as compared to when TMS was applied at −500 and −300 ms. Furthermore, in the go/no‐go RT task, MEP amplitude was significantly smaller at −50 ms and at the go‐signal compared to −150 ms (p‐values < 0.002).

FIGURE 4: Motor‐evoked potential (MEP) integrated area as a function of TMS stimulation time for each reaction time (RT) task (green = simple RT; blue = choice RT; red = go/no‐go RT). Top panel shows fixed effect means and 95% CI for each task. Bottom panel shows fixed effects means and 95% CI with large markers and error bars in separate panels for each task. Individual subject estimated mean MEP amplitudes are shown as small red markers, and violin plots represent the distribution of raw data. Note that means are back‐transformed from the log scale. Asterisk (*) indicates significant difference between tasks at each time point. Dagger (†) indicates significant difference between time points (gray = all tasks, red = go/no‐go only).

When comparing between RT tasks at each TMS time point, MEP amplitude for the choice RT task was significantly larger than that for the go/no‐go RT task (p < 0.001) when TMS was applied either 50 ms before or coincident with the go‐signal. Similarly, MEP amplitude for the choice RT task was significantly larger than that for the simple RT task when TMS was applied coincident with the go‐signal (p = 0.010) but did not reach significance for the 50‐ms time point (p = 0.091).

MEP Suppression as a Function of RT

Change in MEP amplitude was analyzed as a function of RT when TMS was applied at the go‐signal to determine if RT was predictive of the magnitude of MEP suppression. Results for all three RT tasks are presented in Figure 5 with simple RT, choice RT, and go/no‐go RT presented in separate panels. There was a significant effect of RT on the change in MEP amplitude, F(1,542.7) = 4.362, p = 0.037; however, there was no effect of RT task (p = 0.984) and no interaction between the predictors (p = 0.809).

FIGURE 5: Change in motor‐evoked potential (MEP) amplitude at the go‐signal (0 ms) as compared to the warning signal (−500 ms) for each reaction time (RT) task (green = simple RT; blue = choice RT; red = go/no‐go RT). Note that negative values represent MEP suppression as compared to the −500‐ms time point. Thick black lines show an overall positive relationship for all three RT tasks, whereby shortest RTs are associated with largest suppression. Light gray lines show the relationship for each participant, and dots represent individual trials.

Background EMG

Analysis of background EMG ratio resulted in a significant main effect of RT task, F(2,4785.8) = 5.339, p = 0.005. Post hoc analysis indicated that the background EMG ratio was significantly larger (p = 0.003) in the simple RT task (mean = 1.030, 95% CI [1.003, 1.060]) as compared to the choice RT task (mean = 0.982, 95% CI [0.956, 1.010]), whereas the other pairwise comparisons were not significant (go/no‐go mean = 1.008, 95% CI [0.982, 1.030]; p‐values > 0.152). There was no effect of TMS time (p = 0.225) and no interaction between the predictors (p = 0.851).