Common and unique components of response inhibition revealed by fMRI
Introduction
Withholding inappropriate responses is a hallmark of human behavior. This can be observed in many individual tasks in which the tendency to make an automatic or natural response must be suppressed in order to make an appropriate but unnatural response. Much research has been concerned with cases in which conflicting motor responses of this sort are engaged (e.g., Dagenbach and Carr, 1994, DeJong et al., 1995, Anderson and Spellman, 1995, Kornblum et al., 1990). It is often argued that resolving cases of motor conflict requires the engagement of inhibitory processes that dampen the tendency to make the inappropriate response in favor of the appropriate one (e.g., Kornblum and Requin, 1995, Logan, 1985, Logan and Cowan, 1984, Lowe, 1979). The issue we address is, are these inhibitory processes all of a sort? That is, when one has to inhibit an inappropriate motor response in one situation, are the same mechanisms recruited as when one has to inhibit a response in another situation?
We engaged this issue by conducting a simple experiment, collecting behavioral and fMRI data: All participants were presented with three tasks in which avoiding an inappropriate response was required. The three tasks varied in their specific demands and perceptual components, but they shared a common need to inhibit inappropriate responses. One was a stimulus–response compatibility task (SRC), which, on some trials, required overcoming a compatible stimulus–response mapping in order to respond appropriately to each stimulus. For example, upon viewing an arrow pointing to the left, one might be required to either press a key with the left hand (compatible mapping) or the right hand (incompatible mapping). We compared responses in the incompatible mapping condition to those in the compatible one to isolate interference-resolution processes (including response inhibition) involved in this task. A second task involved the go/no-go paradigm (GNG), in which two types of stimuli were presented, one requiring a response and the other requiring the withholding of a response. Participants were instructed to respond as quickly as possible to each target stimulus – for example, the letter X– and withhold a response to a non-target stimulus (e.g., the letter Y). We compared “no-go” trials, which required response inhibition, with “go” trials in which the response was executed. The third situation was a flanker task in which a response had to be made to a central stimulus while ignoring flanking stimuli (Eriksen and Eriksen, 1974). For example, a centrally presented color patch (blue or yellow) was mapped to a right-hand response. The two flanking color patches were mapped to the same response (congruent) or a different response (incongruent) as the target. In this task, we compared trials in which the flanking stimuli were mapped onto incongruent vs. congruent responses.
The type of response competition in this flanker task may be similar to that in the SRC task, but the differences may be important as well: incongruent flankers interfere with incorrect responses through a newly learned color–response association rather than an inherent compatibility effect, as in SRC tasks (Kornblum et al., 1990, Zhang et al., 1999). Also, spatial selective attention may be profitably used in the flanker task, as flankers are spatially distinct from the target, but they are not in the SRC task. Similarly, the inhibition required in the GNG task may be different from that of the flanker and SRC tasks as it involves the withholding of a prepotent response, rather the production of an alternate response.
The question we asked of these three cases was this: Is the same signature of brain activation obtained across tasks? Alternatively, are there distinct patterns of brain activation that are tied to individual situations in which inhibition is required? In either case, we were also interested in whether the pattern or patterns of brain activation would correlate with the corresponding behavioral measures of inhibition.
The question at hand is a pressing one because there is a good deal of controversy in the literature about whether there are common mechanisms of inhibition at work in these and other similar tasks. On the one hand, some previous research with response interference paradigms has shown a relatively cohesive set of areas that are involved in these inhibitory tasks. The anterior cingulate, dorsolateral prefrontal, inferior prefrontal, premotor, and parietal cortices have all been implicated in SRC tasks (Dassonville et al., 2001, Iacoboni et al., 1996, Merriam et al., 2001, Peterson et al., 2002, Schumacher and D'Esposito, 2002), GNG tasks (Casey et al., 1997, Klingberg and Roland, 1997, Konishi et al., 1999, Liddle et al., 2001, Menon et al., 2001, Rubia et al., 2001), and flanker tasks (Bunge et al., 2002a, Bunge et al., 2002b, Casey et al., 2000, Hazeltine et al., 2000). On the other hand, when examined behaviorally, the correlations in performance between one inhibitory task and another, while sometimes significant, generally explain little of the individual variability in these tasks (Burgess et al., 1998, Duncan et al., 1997, Fan et al., 2003, Miyake et al., 2000).
One issue unresolved in previous research is that studies that putatively target “inhibitory processes” generally use a single task and rely on subtraction methods or parametric variation to isolate the “inhibitory” (i.e., interference-resolution) processes of interest. For example, the Stroop task, in which participants must name the ink color of a color word (e.g., “RED” printed in blue ink) while ignoring the word form, is perhaps the classic example of an “inhibitory” task. However, behavioral research has shown that the Stroop task is likely to involve multiple types of interference, with multiple ways of resolving them (Kornblum et al., 1990). Irrelevant information may be inhibited at perceptual, semantic, or response-selection stages of processing. Thus, activations in these tasks could arise from several factors that differ between “inhibition” and control conditions, including enhanced attention to the task in general, increased demand for divided attention, enhanced long-term memory demands when maintaining the less automatic task-set in the inhibition task, and other factors. Through our selection of tasks, we attempt to reduce the ambiguity in interpretation resulting from these multiple types of interference.
Two features of our design help circumvent the abovementioned problems. First, by studying three tasks that share the common requirement for response inhibition, we can make stronger conclusions about areas that are activated in all three tasks. Second, we chose tasks such that response inhibition/selection is ostensibly the only type of inhibition common to all three tasks.
In the GNG task, one has to inhibit the prepotent tendency to execute a response. This sort of inhibition may occur only at the response-selection or execution stages (Rubia et al., 2001), as there is only a single imperative stimulus that must be attended at any given time, and no prior learning builds an association between the “no-go” stimulus and either the “go” stimulus or a “go” response. Thus, there is little stimulus or response overlap that could lead to other forms of interference (Kornblum et al., 1990, Zhang et al., 1999).
In the flanker task, flanking stimuli are spatially distinct from the target, and incompatible flankers overlap with the target response but not with the target stimulus (Kornblum et al., 1990). Thus, inhibition may occur at two levels. First, one may inhibit the perceptual processing of the competing stimuli. However, should this type of inhibition fail, the competing stimuli may activate a competing response, which must be inhibited during response selection (van Veen et al., 2001). As this is the only task with spatially distinct stimuli that could be perceptually filtered (Broadbent, 1977, Mangun, 1995), regions uniquely activated by the flanker task may map most directly to perceptual selection.
In the incompatible trials of the SRC task, a prepotent tendency is developed through overlap between the automatic response elicited by the stimulus (e.g., left arrow) and the incorrect response (left button), competing with the correct response (right button) (Kornblum et al., 1990). Since interference resolution cannot occur during perceptual selection, it is likely to occur at the stage of response selection. However, it is also possible that rules directly compete, and priming of the incorrect rule or task set produces interference (Allport et al., 1994, Monsell et al., 2000, Rubenstein et al., 2001).
Therefore, although each task may recruit unique processes (response execution inhibition in GNG, perceptual inhibition in flanker, and stimulus–response mapping inhibition or rule selection in SRC), what these tasks share in common is inhibition at the level of response selection (e.g., Nee et al., 2004). In addition, by studying the three tasks in the same participants, we circumvented problems that arise when comparing across studies that use different individual brains, different standard brain spaces, and different analysis methods.
There are a few previous reports in the literature that compare different inhibitory tasks within the same individuals, and they do report common regions of activation due to inhibitory processes (Fan et al., 2003, Peterson et al., 2002). These studies report low behavioral correlations among inhibition tasks, and they concluded that overlapping regions alone do not provide strong support for the existence of common mechanisms underlying inhibitory tasks. However, these studies did not correlate behavioral interference measures with regions of activation (but see Bunge et al., 2002a, Bunge et al., 2002b). Additional evidence about relationships between brain activity and behavioral performance could strengthen the case for common brain regions for interference resolution and provide leverage in interpreting the functional significance of overlapping activations. Beyond replicating and extending previous work, another critical issue is an examination of the correlations between activation in these carefully isolated “common” regions with behavioral measures of inhibitory control. Our experiment was designed to address these issues, and in our analysis we specifically searched for regions that showed both activation and correlations with performance in each inhibitory task. In addition, we used connectivity analyses on the pattern of individual differences (e.g., Lin et al., 2003) to ask (a) whether activated regions are organized into coherent, distributed networks of brain regions showing similar patterns of task-related individual differences, and (b) whether these networks are the same across inhibitory tasks.
Section snippets
Participants
Fourteen undergraduate students (ages 18–25) from the University of Michigan were recruited through advertisements placed in the campus paper and flyers posted in campus buildings. All participants completed a self-report health screen for neurological and psychiatric diagnoses, as well as drug or alcohol abuse. They signed informed consent forms approved by the University Institutional Review Board and were compensated up to US$40 for their participation.
Behavioral tasks
There were 6 runs in the scanning
Behavioral results
Behavioral analyses revealed significant interference caused by the active inhibition component of each task. In all three tasks (GNG, flanker, and SRC), the behavioral data were analyzed on a block level, comparing accuracy and response times for the low-conflict blocks with those in the high-conflict blocks (i.e., high “no-go”, congruent, and compatible blocks compared to low “no-go”, incongruent, and incompatible blocks).
In the GNG task, the false alarm rate (FAR) on no-go trials was used as
Discussion
This study demonstrates that diverse inhibitory tasks – one that required withholding of a response (GNG), one that required inhibiting encoding and/or responses (flanker), and one that required re-mapping stimulus–response associations (SRC) – shared substantial overlap in neural activations in the same participants. A core set of commonly activated regions, including bilateral insula, anterior PFC, right DLPFC, and right SFS, were also correlated with one another across participants in each
Conclusions
We found evidence for common activation of a number of brain structures in three interference tasks that share a requirement for response selection. The regions include bilateral insula/frontal operculum, caudate, lateral prefrontal cortices, anterior cingulate, and right premotor and parietal cortices. A subset of these (most consistently the insula) were positively correlated with poorer behavioral performance in each task. The common locations of performance-correlated activations suggests
References (67)
- et al.
Factor analysis of four measures of prefrontal lobe functioning
Arch. Clin. Neuropsychol.
(1998) - et al.
Neural mechanisms of transient and sustained cognitive control during task switching
Neuron
(2003) - et al.
Immature frontal lobe contributions to cognitive control in children: evidence from fMRI
Neuron
(2002) - et al.
Dissociable contributions of prefrontal and parietal cortices to response selection
NeuroImage
(2002) - et al.
A common network of functional areas for attention and eye movements
Neuron
(1998) - et al.
The effect of stimulus–response compatibility on cortical motor activation
NeuroImage
(2001) - et al.
Does word training increase or decrease interference in a Stroop task?
J. Exp. Child Psychol.
(1979) - et al.
Cognitive and brain consequences of conflict
NeuroImage
(2003) - et al.
Thresholding of statistical maps in functional neuroimaging using the false discovery rate
NeuroImage
(2002) - et al.
Thalamic relay functions and their role in corticocortical communication: generalizations from the visual system
Neuron
(2002)
Interference between two concurrent tasks is associated with activation of overlapping fields in the cortex
Brain Res., Cogn. Brain Res.
Multivariate analysis of neuronal interactions in the generalized partial least squares framework: simulations and empirical studies
NeuroImage
Efficiency, power, and entropy in event-related fMRI with multiple trial types: Part II. Design of experiments
NeuroImage
Detection power, estimation efficiency, and predictability in event-related fMRI
NeuroImage
Executive control of thought and action
Acta Psychol.
Stimulus–response incompatibility activates cortex proximate to three eye fields
NeuroImage
The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: a latent variable analysis
Cogn. Psychol.
An event-related functional MRI study comparing interference effects in the Simon and Stroop tasks
Cogn. Brain Res.
Mapping motor inhibition: conjunctive brain activations across different versions of go/no-go and stop tasks
NeuroImage
Separate modifiability, mental modules, and the use of pure and composite measures to reveal them
Acta Psychol. (Amst.)
Anterior cingulate cortex, conflict monitoring, and levels of processing
NeuroImage
Optimization of experimental design in fMRI: a general framework using a genetic algorithm
NeuroImage
Valence, gender, and lateralization of functional brain anatomy in emotion: a meta-analysis of findings from neuroimaging
NeuroImage
Neuroimaging studies of shifting attention: a meta-analysis
NeuroImage
Accounting for nonlinear BOLD effects in fMRI: parameter estimates and a model for prediction in rapid event-related studies
NeuroImage
A parallel distributed processing model of stimulus–stimulus and stimulus–response compatibility
Cogn. Psychol.
Parallel organization of functionally segregated circuits linking basal ganglia and cortex
Annu. Rev. Neurosci.
Shifting intentional set-exploring the dynamic control of tasks
Atten. Perform. Xv
On the status of inhibitory mechanisms in cognition: memory retrieval as a model case
Psychol. Rev.
When is a conjunction not a conjunction
The hidden preattentive processes
Am. Psychol.
The ecological validity of tests of executive function
J. Int. Neuropsychol. Soc.
A developmental functional MRI study of prefrontal activation during performance of a go-no-go task
J. Cogn. Neurosci.
Cited by (395)
The effect of task complexity on the neural network for response inhibition: An ALE meta-analysis
2024, Neuroscience and Biobehavioral ReviewsMethylphenidate Enhances Spontaneous Fluctuations in Reward and Cognitive Control Networks in Children With Attention-Deficit/Hyperactivity Disorder
2023, Biological Psychiatry: Cognitive Neuroscience and NeuroimagingEffects of neuromodulation on cognitive and emotional responses to psychosocial stressors in healthy humans
2023, Neurobiology of Stress