Elsevier

Brain Research

Volume 1189, 16 January 2008, Pages 127-134
Brain Research

Research Report
ERP—Correlates of response selection in a response conflict paradigm

https://doi.org/10.1016/j.brainres.2007.10.076Get rights and content

Abstract

Neuroimaging and electrophysiological studies suggest that the anterior cingulate cortex (ACC) is involved in the cognitive control of response related action. A frontocentral negative ERP-component, the N2, which probably originates from the ACC, is usually enhanced in conflict-trials that demand an unexpected response. We here used stepped adjustment of response expectation in a response-cueing task, and measured how the N2 varied with global and local cue validity. Results showed that, irrespective of the current cue validity, response times, error rates, and the frontocentral components P2, N2 and P3 increased in unexpected trials. Nevertheless, a N2 was also seen in expected trials, and its latency correlated positively with reaction times, indicating that this potential does not express response conflict only. In line with roles suggested for the ACC, we here propose that the N2 is related to the process of response selection which influences subsequent processing stages reflected in the P3. Unexpected revisions of response programs enhance and delay the N2.

Introduction

Neuroimaging studies suggest that the anterior cingulate cortex (ACC) is involved in response conflict monitoring (see Botvinick et al., 2004 for a review). Conflict arises when incompatible response tendencies are simultaneously active. The frontocentral N2, an ERP component attributed to the ACC (e.g. van Veen and Carter, 2002), is especially prominent in rare conflict trials. When conflict and no-conflict trials are equiprobable, it is smaller (Braver et al., 2001), suggesting that the N2 is sensitive to response probability. Would rare compatible trials also elicit a N2?

While the ACC's role in conflict processing is widely accepted (see Botvinick et al., 2004 for a review), the ACC, in cooperation with other brain areas, also supports other functions. It is the crucial brain area involved in selecting and coupling perceptual information to motor action (Badgaiyan and Posner, 1998, Holroyd and Coles, 2002, Posner et al., 1988). Recent studies suggest that ACC takes part in response selection, i.e. the cognitive process of assigning a specific response to a specific stimulus category (Isomura et al., 2003, Paus, 2001, Picard and Strick, 1996, Turken and Swick, 1999). One line of evidence for the response selection account of the ACC stems from neuropsychological investigations. Turken and Swick (1999) evaluated a patient who had suffered a focal lesion affecting part of the right ACC. The authors tested the patient in a variety of executive control tasks. Reasoning that performance should be impaired regardless of the response modality if the ACC was responsible for general executive control, they requested both manual and vocal responses. Their results showed that performance was only impaired in manual response trials. Moreover, the impairment was most evident when response selection requirements were most demanding, that is in trials with a high level of conflict. A role for the ACC in response selection is also suggested by neurophysiological data. Reviewing the functional properties of the primate medial wall, Picard and Strick (1996) reported that the rostral zone of the ACC participated in tasks requiring movement selection and willful generation of many different types of motor behavior. Paus (2001) suggested that the lateral prefrontal cortex (PFC) computes and maintains information necessary for choosing the appropriate response, whereas the ACC facilitates implementation of action. Its overlapping motor, cognitive and arousal domains place the ACC in a unique position to translate intentions to action. More evidence for the response selection account of the ACC comes from single unit recordings in monkeys. Isomura et al. (2003) found that rostral cingulate motor areas (a homologue of a human ACC) participate in appropriate action according to an intention whereas dorsal and ventral ACC may be involved in motor preparation and execution. Consequently, the ACC responded only to stimulus classes that were associated with a particular response indicating an integrative role of the ACC in coupling of stimuli with a particular action. Finally, a recent fMRI study supported by computer simulations (Roelofs et al., 2006) challenged the conflict hypothesis of ACC in favour of the selection-for-action hypothesis. Roelofs and co-workers observed enhanced ACC activation in the absence of response conflict in a Stroop-like task. According to the regulative hypothesis of the ACC proposed by those authors, both Stroop facilitation (congruent vs. neutral) and Stroop interference (incongruent vs. neutral) have a common source and arise during response selection by selectively enhancing the activation of the correct response until a selection threshold is exceeded.

Finally, evidence relating the N2 to response selection also comes from the ERP literature. Ritter et al., 1982, Ritter et al., 1983 first related the N2 to stimulus classification in choice tasks, which is close to our definition of response selection. Hohnsbein et al. (1998) could show that in normal young subjects who performed a speeded choice reaction task, the N2 (then named N-CR) and the subsequent parietal P3 (P-CR) were delayed – while the RTs were not – in subjects who committed many errors compared to those who committed few errors. This suggests that response selection timing is reflected in N2 (and P-CR) latency. More recently, Di Russo et al. (2006) contrasted ERPs in Discriminative and Simple-Response Task between fencers and controls. The fencers responded significantly faster and had 30 ms shorter N2 — latencies than controls.

In the present study we ask whether the N2 can be flexibly adapted to changes of expected conflict. Moreover, we want to assess its relation to response selection by observing the N2 and its relation to response time irrespective of the presence of conflict. To this aim we recorded ERPs in a response-cueing task. Participants responded differentially to two target letters that were preceded by cues. Over three blocks of trials, cue validity decreased from 80 to 50 to 20%; the subjects were not informed about this change. The aim was to investigate whether the changes in cue validity to which the participants adapt would modulate the conflict-related frontocentral N2. If conflict was under strategic control, the N2 difference should disappear when compatible and incompatible trials are equiprobable. When incompatible trials are more frequent than compatible ones in the 20% validity condition, conflict should arise in compatible rather than incompatible trials, instigating a reversed N2 effect. Finally, if the N2 was involved in response selection, its latency should correlate positively with response times in all conditions, regardless of conflict.

Section snippets

Behavioral data

Mean reaction times (RT) and error rates (ER) are depicted in Fig. 1.

For the analysis of response times, error trials (4%) and outliers (below 100 ms or above 1000 ms; 0.9%) were discarded. A repeated-measures ANOVA with the factors Validity and Compatibility revealed no significant main effect of the factor Validity (F < 1) and an effect of Compatibility that narrowly missed the 5% criterion (F(1,13) = 4.58, P = 0.052). However, the two factors interacted significantly (F(2,26) = 6.38, P < 0.01). When

Behavioral data

Our behavioral data show that a decrease in global cue validity affected the impact of cue-target compatibility. A high degree of conflict was expressed in high RTs and error costs for the rare incompatible trials in the first block. In the second block, where compatible and incompatible trials were equiprobable, the effect was abolished, corroborating previous findings (Braver et al., 2001). In the third block, where the incorrectly cued targets were common at 80%, conflict-induced costs

Conclusions

Our results indicate that the N2 varies with changes in response expectancy. We propose that this potential is related to a general process of response selection, which is intensified and prolonged in conflict trials that require revision of the prepared response plan.

Participants

14 healthy, right-handed subjects participated (6 males). Their age ranged from 21 to 35 years (mean 26.6); all had normal or corrected-to-normal visual acuity.

Task, apparatus and procedure

The stimuli and an example of two trials are depicted in Fig. 5.

Participants were required to respond to the centrally presented target, an ‘O’ or an ‘X’ (3 cm × 3 cm). Half of the participants used the left index finger for the letter O and the right index finger for the X; the assignment was reversed for the other half. The response keys

References (23)

  • C.B. Holroyd et al.

    The neural basis of human error processing: reinforcement learning, dopamine, and the error-related negativity

    Psychol. Rev.

    (2002)
  • Cited by (177)

    View all citing articles on Scopus
    View full text