Cost-Benefit Analysis and Conflict Monitoring: the Anterior Cingulate Cortex
The anterior cingulate cortex (ACC) is often activated during tasks with multiple conflicting responses. One theory is that the ACC itself detects conflict, and then up-regulates attention or "cognitive control" to improve the chances that a correct response is chosen. Another theory is that strong activity within dlPFC signals this heightened conflict, and the ACC "kicks in" as an additional resource to help mitigate competition.
However, recent evidence from Milham & Banich suggests that ACC activation may not be limited to situations involving conflict, but may instead be a functionally diverse area where the more anterior or rostral part is involved in error-related processing and only the posterior portion is activated by conflict.
To demonstrate this, the authors used a version of the Stroop task in which the to-be-named color was either the color in which a word was written, or whether it surrounded the word (i.e., the word was superimposed on a block of that color). As in the classic Stroop, the words could also be congruent color words (i.e., RED), incongruent color words (i.e., GREEN) or neutral non-color words (i.e., LOT) (note that all these examples demonstrate the condition where the word was written in that color; in the "surround" condition, all the words were written in grey ink on top of a block of color). Using this design, the authors calculated the percent facilitation (i.e., how much faster subjects were to respond to congruent than neutral trials) and percent interference (i.e., how much slower subjects were to respond to incongruent than neutral trials) for both the classic and "surround" conditions. 18 right-handed subjects completed this task inside a 1.5 T fMRI machine.
The results showed that RT interference was twice as strong for the classic than the surround condition, suggesting that subjects had a particularly difficult time naming the visible color when it was the same color used to write the word. Disjunction analyses showed that the anterior portion of the rostral ACC was specifically sensitive to interference (but not facilitation) whereas the posterior rostral zone of ACC was sensitive to interference and facilitation. Surprisingly, no regions in frontal or parietal cortex showed differential activation between the classic and surround conditions except for anterior inferior prefrontal cortex, which was more highly activated in the classic than the surround condition (but this was not reported as significant).
The authors conclude that they identified two functionally distinct subregions of ACC: one that is sensitive to response conflict (posterior rostral ACC; prACC) and one that is sensitive to error-related processing (anterior rostral ACC; arACC. They suggest that prACC may be involved in "response selection, facilitating infrequent or novel responses, and inhibiting prepotent responses" whereas arACC may be involved in conflict monitoring proper.
Of course, it is possible to suggest that even prACC is involved in conflict monitoring, but that conflict includes not only response conflict but also conflict related to which aspect of a stimulus to process, or "stimulus evaluation."
In a related review article, Botvinick, Cohen & Carter discuss a variety of interpretations of the function of ACC. They also make the distinction between prACC and arACC in the context of errors: some work shows that prACC responds both during errors and during high-conflict settings, whereas arACC responds preferentially to error trials. Finally, prACC may be the source of the error-related negativity signal detected with ERP in response to errors.
Botvinick et al. conclude their review with a fascinating hypothesis: data from gambling tasks suggests that the conflict monitoring subserved by ACC may be sensitive to rewards and response efforts - in other words, calculating some sort of cost/benefit analysis. In this view, ACC is a "action-outcome evaluator."
Related Posts (primarily related to ACC):
Imaging Lapses of Attention
Selection and Updating Frequency in the Attentional Blink
Reversing Time: Temporal Illusions
Developmental Change in the Neural Mechanisms of Risk Perception
However, recent evidence from Milham & Banich suggests that ACC activation may not be limited to situations involving conflict, but may instead be a functionally diverse area where the more anterior or rostral part is involved in error-related processing and only the posterior portion is activated by conflict.
To demonstrate this, the authors used a version of the Stroop task in which the to-be-named color was either the color in which a word was written, or whether it surrounded the word (i.e., the word was superimposed on a block of that color). As in the classic Stroop, the words could also be congruent color words (i.e., RED), incongruent color words (i.e., GREEN) or neutral non-color words (i.e., LOT) (note that all these examples demonstrate the condition where the word was written in that color; in the "surround" condition, all the words were written in grey ink on top of a block of color). Using this design, the authors calculated the percent facilitation (i.e., how much faster subjects were to respond to congruent than neutral trials) and percent interference (i.e., how much slower subjects were to respond to incongruent than neutral trials) for both the classic and "surround" conditions. 18 right-handed subjects completed this task inside a 1.5 T fMRI machine.
The results showed that RT interference was twice as strong for the classic than the surround condition, suggesting that subjects had a particularly difficult time naming the visible color when it was the same color used to write the word. Disjunction analyses showed that the anterior portion of the rostral ACC was specifically sensitive to interference (but not facilitation) whereas the posterior rostral zone of ACC was sensitive to interference and facilitation. Surprisingly, no regions in frontal or parietal cortex showed differential activation between the classic and surround conditions except for anterior inferior prefrontal cortex, which was more highly activated in the classic than the surround condition (but this was not reported as significant).
The authors conclude that they identified two functionally distinct subregions of ACC: one that is sensitive to response conflict (posterior rostral ACC; prACC) and one that is sensitive to error-related processing (anterior rostral ACC; arACC. They suggest that prACC may be involved in "response selection, facilitating infrequent or novel responses, and inhibiting prepotent responses" whereas arACC may be involved in conflict monitoring proper.
Of course, it is possible to suggest that even prACC is involved in conflict monitoring, but that conflict includes not only response conflict but also conflict related to which aspect of a stimulus to process, or "stimulus evaluation."
In a related review article, Botvinick, Cohen & Carter discuss a variety of interpretations of the function of ACC. They also make the distinction between prACC and arACC in the context of errors: some work shows that prACC responds both during errors and during high-conflict settings, whereas arACC responds preferentially to error trials. Finally, prACC may be the source of the error-related negativity signal detected with ERP in response to errors.
Botvinick et al. conclude their review with a fascinating hypothesis: data from gambling tasks suggests that the conflict monitoring subserved by ACC may be sensitive to rewards and response efforts - in other words, calculating some sort of cost/benefit analysis. In this view, ACC is a "action-outcome evaluator."
Related Posts (primarily related to ACC):
Imaging Lapses of Attention
Selection and Updating Frequency in the Attentional Blink
Reversing Time: Temporal Illusions
Developmental Change in the Neural Mechanisms of Risk Perception
posted by Chris Chatham at 8:33:00 AM. Now go to the current Developing Intelligence blog!
0 Comments:
Post a Comment
<< Home