The relative involvement of anterior cingulate and prefrontal cortex in attentional control depends on nature of conflict

Abstract

While numerous studies have implicated both anterior cingulate and prefrontal cortex in attentional control, the nature of their involvement remains a source of debate. Here we determine the extent to which their relative involvement in attentional control depends upon the levels of processing at which the conflict occurs (e.g., response, non-response). Using a combination of blocked and rapid presentation event-related functional magnetic resonance imaging techniques, we compared neural activity during incongruent Stroop trial types that produce conflict at different levels of processing. Our data suggest that the involvement of anterior cingulate and right prefrontal cortex in attentional control is primarily limited to situations of response conflict, while the involvement of left prefrontal cortex extends to the occurrence of conflict at non-response levels.

Introduction

Although generally agreed that anterior regions of the attentional network are responsible for top-down control of activity within the network, the exact regions exerting such control remain a source of debate. While several models assume attentional control is supported by a single structure such as the anterior cingulate cortex (ACC) or dorsolateral prefrontal cortex (PFC) [17], [22], [25], [20], recent studies have suggested the involvement of both regions [10], possibly possessing complementary roles [1], [2], [14]. The emerging perspective from these latter studies is that the ACC may be involved in evaluatory processes, monitoring for the occurrence of conflict within the attentional network [14], [7], [3]. In contrast, the dorsolateral (BA 9 and BA 46) and posterior inferior PFC (BA 44) may be involved in maintaining an attentional set by modulating activity within posterior processing systems and facilitating selection of task relevant representations within working memory [1] (for a similar perspective see Ref. [14]).

Here we consider the extent to which the relative involvement of the ACC and PFC (dorsolateral and posterior inferior) depends upon the level(s) of processing at which the need for attentional control arises. While most studies of attentional control examine the involvement of the ACC and PFC by manipulating whether or not conflict can occur during task performance, we took the approach of varying the stages of processing at which conflict can occur. This was accomplished using the Stroop task, a paradigm in which competition has been shown to occur at multiple levels of performance (i.e., both response and non-response levels) [15].

In the Stroop task, participants identify the ink color in which a word is printed while ignoring its identity. Although seemingly uncomplicated, the word itself is a potentially strong source of interference to color identification. Engagement of the systems dedicated to processing words (e.g., orthography to phonology, orthography to semantics) is relatively automatic, requiring little or no attention [15]. Thus representations (e.g., semantic or phonological) arising from the word processing systems can compete for processing resources. Such competition is especially heightened when the word names a color. In this case, each processing system (i.e., word system, color system) can lead to the activation of not only the semantic and phonological representations associated with a particular color, but also associated responses. Inadvertent contributions from the word processing system are evident, as words incongruent with the ink color degrade performance whereas those congruent facilitate performance compared to neutral trials. Attentional selection is required to ensure that performance is based upon the task-relevant processing of the ink’s color rather than the task-irrelevant color associated with the word’s identity, a function typically attributed to both the PFC (dorsolateral and posterior inferior) and ACC [2], [11], [5], [19].

We reasoned that if the ACC’s involvement in attentional control is mainly limited to situations in which competition is present at the response level, then its activity during the attentionally demanding incongruent condition should be disproportionately influenced by whether or not response competition can occur. To test this idea, we varied whether an incongruent trial could produce interference at both the response and non-response levels, or just the non-response levels. More specifically, we manipulated whether the incongruent color word named a color in the response set (an eligible response) or a color outside of the response set (an ineligible response). Incongruent response-eligible words can be potentially strong sources of interference, due to their ability to introduce competing phonological and semantic representations related to color, as well as competing representations of the response output. In contrast, incongruent-ineligible words can only produce interference at the level of semantics and phonology, since the word does not name a potential response. To examine the degree of interference engendered by both incongruent ineligible and incongruent eligible trials, we compared activation to that which occurs during neutral trials, in which the word contains no color information or associated responses (e.g., LOT).

It is important to note that the inclusion of neutral trials allowed us to determine the extent to which a region’s involvement in attentional control is dependent upon the presence of conflict at the response level. Relative to neutral trials, incongruent-eligible trials can produce conflict at both response and non-response levels. In contrast, incongruent-ineligible trials only differ from neutral trials in their ability to generate conflict at non-response levels, as neither incongruent-ineligible words nor neutral words are associated with an eligible response. Therefore, if activity within a brain region increases for incongruent trials that name an eligible response relative to neutral trials but not for incongruent trials that name an ineligible response, then that region’s involvement in attentional control is primarily limited to the presence of conflict at the response level. On the other hand, if a region shows increased activity for both types of incongruent trials, even when the color word does not name a possible response, then its involvement in attentional control is clearly not limited to situations of conflict at the level of response.

Given that the ACC’s involvement in attentional control appears to be dependent upon the presence of response competition [1], [18], [25], we expected to find little to no activity during incongruent-ineligible trials when compared to neutral trials, but increased activity on incongruent-eligible trials. In contrast, dorsolateral and posterior inferior prefrontal cortices, which have been shown to be involved in attentional control when conflict arises at non-response levels trials [1], [14], should be activated for both incongruent trial types when compared to neutral.