Filling the gaps: Cognitive control as a critical lens for understanding mechanisms of value-based decision-making

Abstract

Research into value-based decision making has made tremendous progress in identifying behavioral and neural correlates of choice value. However, these correlates have been primarily viewed through a field-specific lens, focusing on how they contribute to the evaluation and selection between options to arrive at a choice. Here, we reveal blind-spots resulting from this limited perspective, and how they can be filled in through taking the perspective of cognitive control. We highlight three particular insights that this perspective offers: (1) a view towards the goal-relevance of one’s options and their features; (2) a view of decision-making correlates as a proxy for monitoring to determine control adjustments; (3) a view of those correlates as a proxy for monitoring that extends temporally and hierarchically beyond the immediate choice task. We show how adopting these complementary perspectives offers new insight into the determinants of both decisions and control; alternative interpretations for common findings in the neuroeconomic literature; and fruitful directions for future research.

Advances in modeling learning and decision-making in neuroscience

“…Signatures of this evidence accumulation process have been identified throughout the brain [131,132], with some notable differentiation between neural circuits that appear to track the accumulation of information about which options are the most valuable (e.g., in orbital and/or ventromedial PFC [128,133,134]); which response is the best overall (e.g., in dorsomedial prefrontal or lateral intraparietal regions [131,135,136]); and what motor action to implement (e.g., in premotor and/or motor cortices [137][138][139]). However, as we discuss next, the interpretation of these signals as indexing evidence accumulation per se, rather than a covariate thereof, is a matter of significant debate [140].…”

“…Work has shown that a person can evaluate these same metacognitive signals of confidence/uncertainty while making their choice (e.g., based on the overall strength of evidence), and can make online adjustments to their choice strategy accordingly (e.g., increasing their decision threshold when less confident) [143,145,149,150]. Metacognitive evaluations thus offer a prominent alternative explanation for putative neural correlates of evidence accumulation, including estimates of the decision variable itself [127,140,[151][152][153][154].…”

“…Another area for future research is to better understand the representations that these computations operate over [18,140,248]. We focused earlier on state and action representation, where many questions remain open, but another (arguably even more fundamental) question pertains to how people represent outcomes, in particular what counts as a reward.…”