Learning from errors: error-related neural activity predicts improvements in future inhibitory control performance

Abstract

Failure to adapt performance following an error is a debilitating symptom of many neurological and psychiatric conditions. Healthy individuals readily adapt their behavior in response to an error, an ability thought to be subserved by the posterior medial frontal cortex (pMFC). However, it remains unclear how humans adaptively alter cognitive control behavior when they reencounter situations that were previously failed minutes or days ago. Using functional magnetic resonance imaging, we examined neural activity during a Go/No-go response inhibition task that provided the opportunity for participants to learn from their errors. When they failed to inhibit their response, they were shown the same target stimulus during the next No-go trial, which itself could occur up to 20 trials after its initial presentation. Activity within the pMFC was significantly greater for initial errors that were subsequently corrected than for errors that were repeated later in the display sequence. Moreover, pMFC activity during errors predicted future responses despite a sizeable interval (on average 12 trials) between an error and the next No-go stimulus. Our results indicate that changes in cognitive control performance can be predicted using error-related activity. The increased likelihood of adaptive changes occurring during periods of recent success is consistent with models of error-related activity that argue for the influence of outcome expectancy (Holroyd and Coles, 2002; Brown and Braver, 2005). The findings may also help to explain the diminished error-related neural activity in such clinical conditions as schizophrenia, as well as the propensity for perseverative behavior in these clinical groups.

Figure 1.

Sample displays from the Go/No-go task. A series of letters was displayed at a rate of 1 per second. Participants were required to respond to each of the Go-trials with a button press, and to withhold their response whenever an identical letter was presented on consecutive trials. The figure presents an example of a No-go trial (the repetition of the letter “C”), followed by the presentation of the same stimulus on the next No-go trial (the second repetition of the letter “C”). This stimulus sequence only occurred when participants had unsuccessfully inhibited their response to the first No-go trial.

Figure 2.

Classification of No-go events. No-go trials were categorized based on performance (errors, marked with a cross; and stops, marked with a tick) and by performance on the subsequent No-go trial (relationship indicated by dotted lines). Correction-predicting errors were No-go errors that were followed by correct performance on the subsequent No-go trial (A), whereas error-predicting errors were followed by another No-go error (B, C). Stops (correct response inhibition) were also categorized into those presenting the same stimulus letter (stop “same”) or a different letter (stop “different”) as the previous No-go trial. The second consecutive failure to inhibit for the same letter stimulus was labeled as a “repeat” error, which could also be considered “adaptive” (B) or “maladaptive” (C) depending on performance for the subsequent No-go trial. Note that the No-go trial following a repeat error would always present a different stimulus letter.

Figure 3.

Go trial reaction times for the period following correction- and error-predicting errors. Mean reaction times are presented for the Go trial preceding a No-go error (preNo-go), the failed No-go trial (No-go), and the Go trials that followed the error. Although the number of Go trials following an error ranged from 1 to 22, the range for the analysis was restricted to 16 trials to improve reliability. The later trials (i.e., 17–22) were excluded because some participants made too few errors to obtain a reliable estimate of response speed. Repeat errors represent the second error in a consecutive pair of No-go errors and are distinct from first-presentation errors because they were not predictive of the next No-go trial (which was a different random letter).

Figure 4.

3D rendering of brain regions differentiating corrected from error-predicting errors. Correction-predicting errors, when compared with error-predicting errors, were associated with significantly higher levels of activity in the following: axial view of the error-related cingulate cluster (MNI: x = 2; y = 12; z = 44) (I); the right middle temporal gyrus (x = 48; y = −27; z = −9) (II); the right inferior parietal (x = 53; y = −42; z = 32) and left inferior parietal (x = −54; y = −41; z = 39) gyri (III); the right insula (x = 40; y = 14; z = 5) and left insula (x = −40; y = 13; z = 4) (IV); and the right inferior frontal gyrus (x = 49; y = 39; z = 12) (V).