McGovern, Robert A.; McKhann, Guy M.
One of the hallmarks of human cognition is the ability to link contextual information with behavioral actions. The Gratton effect1 is one well-known example of this ability: humans accelerate behavioral responses to promote efficiency when the current task is similar to a prior task, and slow their responses to promote accuracy when the current task is different from a prior task. The dorsal anterior cingulate cortex (dACC) has been implicated in this process in prior studies demonstrating its role in detecting errors,2 monitoring conflict,3 and integrating monetary reward with behavior.4 The vast majority of these studies have used functional magnetic resonance imaging (fMRI) as the technique of choice while subjects perform a task designed to investigate the cognitive ability in question. Recently, Sheth and colleagues in the Department of Neurosurgery at Massachusetts General Hospital used a combination of intraoperative single neuron microelectrode recordings and fMRI in patients undergoing selective anterior cingulotomy for the treatment of refractory obsessive-compulsive disorder (OCD). Their work, published in Nature, is a seminal investigation into the function of the dACC (Nature. 2012 Jun 24. Doi: 10.1038/nature11239. [Epub ahead of print]).
Subjects in this study performed an interference task preoperatively while undergoing fMRI; intraoperatively while single neurons were being recorded from the dACC; and postoperatively (Figure 1). The task was designed such that the subject viewed 3 numbers displayed in sequence from a set of 4 possible numbers (0,1,2,3). The subject then had to choose the target number (1,2,3) displayed that was different from the other numbers displayed by pushing a button specific to that number (left button for 1, middle for 2, right for 3). Zero was displayed but could not be selected. There are number displays (020) that can be considered low interference (type 0 trial) because the number display is spatially congruent (2 is displayed in middle position which correlates with button to be pressed) and has no potential distractors (0 is not a possible selection). Mid-level, or type 1, interference trials possessed either a spatially incongruent display (200) or relevant distractors (323). High interference, or type 2, trials (233) were both spatially incongruent and had relevant distractors (Figures 1A and 1B).
As prior studies have shown, reaction time (RT) increased as the level of interference increased in a dose-dependent manner (Figure 1C). When comparing type 0 to type 2 trials, fMRI showed an increased level of activity in dACC among other areas (Figure 1D). Single unit recordings from dACC demonstrated that, among 3 distinct groups of neurons, the largest population was composed of neurons that fired preferentially following the cue. This group, considered to be the “cue responsive population,” was investigated in more detail. In these neurons, their firing rate increased in type 2 vs type 0 trials. The authors also found that, in trials where the number of responses was constant but conflict varied, firing rate increased with an increase in conflict. Thus, they suggest that these cue responsive neurons in the dACC are responsible for encoding conflict detection.
Importantly, the investigators found that subjects exhibited the Gratton effect: a tendency for prior trials of a similar type to decrease RT. Thus, regardless of level of interference, RT was shorter when the previous trial type was the same, and conversely, increased when the previous trial type was different. Single unit recordings showed that firing rate increased when the prior trial contained interference, again suggesting that these neurons encode an interference, or conflict signal. Intriguingly, cingulotomy lesions did not affect global performance with subjects showing similar RTs across trial types post-cingulotomy. However, post-cingulotomy subjects showed an abolition of the Gratton effect with similar RTs after a previously similar trial type.
This elegantly designed study provides unique insight into the neurophysiology of the dACC. It demonstrates the basis of a well-known phenomenon, the Gratton effect, in humans using single neuron recordings and fMRI, with subsequent abolition after a lesion relatively specific to the dACC. It also suggests that the dACC is the seat of the Gratton effect, providing a continuously updated readout of cognitive demand such that when conflict arises and cognitive demand increases, the dACC increases firing rate, and slows the response to promote accuracy. When cognitive demand remains stable, the dACC accelerates the response and promotes efficiency. Because global task performance was unaffected by the lesion, the basis for the act of action selection itself likely lies elsewhere, perhaps in the basal ganglia, and raises a new set of questions to be answered. This study highlights the power of using patient based intraoperative research to provide fundamental insights into human cognition.
1. Gratton G, Coles MG, Donchin E. Optimizing the use of information: strategic control of activation of responses. J Exp Psychol Gen. 1992;121(4):480–506.
2. Brown JW, Braver TS. Learned predictions of error likelihood in the anterior cingulate cortex. Science. 2005;307(5712):1118–1121.
3. Botvinick M, Nystrom LE, Fissell K, Carter CS, Cohen JD. Conflict monitoring versus selection-for-action in anterior cingulate cortex. Nature. 1999;402(6758):179–181.
4. Williams ZV, Bush G, Rauch SL, Cosgrove GR, Eskandar EN. Human anterior cingulate neurons and integration of monetary reward with motor responses. Nat Neurosci. 2005;7(12):1370–1375.