Distinct neural processes support post-success and post-error slowing in the stop signal task

Abstract

Executive control requires behavioral adaptation to environmental contingencies. In the stop signal task (SST), participants exhibit slower go trial reaction time (RT) following a stop trial, whether or not they successfully interrupt the motor response. In previous fMRI studies, we demonstrated activation of the right-hemispheric ventrolateral prefrontal cortex, in the area of inferior frontal gyrus, pars opercularis (IFGpo) and anterior insula (AI), during post-error slowing (PES). However, in similar analyses we were not able to identify regional activities during post-success slowing (PSS). Here, we revisited this issue in a larger sample of participants (n=100) each performing the SST for 40 min during fMRI. We replicated IFGpo/AI activation to PES (p≤0.05, FWE corrected). Further, PSS engages decreased activation in a number of cortical regions including the left inferior frontal cortex (IFC; p≤0.05, FWE corrected). We employed Granger causality mapping to identify areas that provide inputs each to the right IFGpo/AI and left IFC, and computed single-trial amplitude (STA) of stop trials of these input regions as well as the STA of post-stop trials of the right IFGpo/AI and left IFC. The STAs of the right inferior precentral sulcus and supplementary motor area (SMA) and right IFGpo/AI were positively correlated and the STAs of the left SMA and left IFC were positively correlated (slope>0, p's≤0.01, one-sample t test), linking regional responses during stop success and error trials to those during PSS and PES. These findings suggest distinct neural mechanisms to support PSS and PES.

Keywords: cognitive control; error processing; fMRI; go/no-go; post-signal slowing.

Figure 1.

The stop signal task (SST) and performance. (A) Behavioral paradigm; (B) A trial sequence to illustrate post-go (pG), post-stop success (pSS), and post-stop error (pSE) go trials; (C) Bar plots to show the extent of post-success slowing (PSS) and post-error slowing (PES) in mean ± S.D. across subjects: *p≤0.001 and **p≤0.0001; (D) A scatter plot to show the positive correlation between PSS and PES across subjects. Each dot represents data of one subject.

Figure 2.

Regional activations to post-error slowing (PES) and post-success slowing (PSS). IFGpo/AI: inferior frontal gyrus, pars opercularis/anterior insula; IFC: inferior frontal cortex. The results of contrast (A) pSEi vs. pSEni and (B) pSSi vs. pSSni, at p≤0.001, uncorrected, shown in axial sections from z=−32 to +64 and neurological orientation: R = right. Color scale reflects voxel T values.

Figure 3.

Brain regions providing Granger causality inputs to (A) right IFGpo/AI, including the thalamus, right insula, right inferior precentral sulcus (IPcS), precuneus (PCu), and right supplementary motor area (SMA); and to (B) left IFC, including the right posterior insula and left SMA.

Figure 4.

Linear correlation of single trial amplitude (STA) between target region – right inferior frontal gyrus, pars opercularis and anterior insula (rIFGpo/AI) – and five input regions – thalamus, right insula, right inferior precentral sulcus (IPcS), precuneus (PCu), and the right supplementary motor area (SMA); as well as between target region – left inferior frontal cortex (lIFC) – and two input regions – insula and the left SMA. The regression lines are aligned at an intercept of zero for visualization of the overall pattern of slopes.