Dissociable neural mechanisms track evidence accumulation for selection of attention versus action

Amitai Shenhav, Mark A. Straccia, Jonathan D. Cohen, Matthew Botvinick

TL;DR

This study reveals how our brains simultaneously process information for choosing what to focus on and what action to take. Using fMRI, researchers found distinct areas in the prefrontal cortex that track reward value versus attentional priority, and even specific sub-regions within the anterior cingulate cortex that encode uncertainty at different decision-making levels. This work bridges the gap between perceptual and value-based decision-making, showing a dynamic, parallel integration of information.

What Problem Does This Paper Address?

Decision-making is often viewed as a sequential process: first, we decide what to pay attention to (e.g., which task or stimulus feature is important), and then we decide what action to take based on that attended information. However, in real life, these processes often happen in parallel. For instance, while you're deciding which email to reply to, you might also be evaluating the salience of other incoming messages. This paper tackles the question of how the brain handles this parallel processing, specifically how it integrates evidence for selecting attentional targets (higher-level decisions) with evidence for selecting motor responses (lower-level decisions), and how reward value influences this integration. Previous research has identified brain regions involved in each aspect separately (e.g., vmPFC for value, dACC for uncertainty in response selection), but how they interact in a parallel decision-making context remained unclear.

Methods

The researchers used functional magnetic resonance imaging (fMRI) to scan participants while they performed a novel value-based perceptual decision-making task. This task involved participants deciding between two dynamic stimulus attributes (motion direction and color) that offered different rewards. Participants had to simultaneously weigh the evidence for each attribute and decide which response to make. The study employed mixed-effects logistic regressions to analyze behavioral data and mixed-effects regressions on fMRI beta estimates to link neural activity to behavior. They also used a whole-brain analysis to explore regions beyond their pre-defined regions of interest (ROIs) and defined ROIs based on previous literature for vmPFC and dACC, with further subdivisions of dACC along its rostrocaudal axis.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Parallel Processing is Key: The brain doesn't strictly separate "what to attend" from "what to do." It juggles these decisions concurrently.
  • Distinct Neural Roles: Different parts of the prefrontal cortex have specialized jobs. vmPFC seems to track the value of information, while dACC tracks attentional priority or how much an attribute influences the choice.
  • Hierarchical Uncertainty: Within the dACC, there's a spatial organization. More "rostral" (front) parts seem to handle uncertainty about what to attend, while more "caudal" (back) parts handle uncertainty about which response to make. This suggests a functional topography for decision-making uncertainty.
  • Reward Shapes Attention: When one attribute is more rewarding, participants (and their brains) tend to prioritize it, even if it's less perceptually salient. This highlights the interplay between value and attention.

Key Results

  • vmPFC: Positively tracked the evidence for the chosen option, with its weighting of attributes proportional to their expected reward. It also showed effects of trial history, encoding reward from the previous trial.
  • dACC: Negatively tracked the evidence for the chosen option, suggesting it might be more involved in signaling uncertainty or conflict.
  • Rostral dACC: Positively encoded the salience of the low-reward attribute, suggesting it tracks uncertainty about which attribute to attend, especially when a tempting alternative arises.
  • Caudal dACC: Was sensitive to uncertainty at the response level (left vs. right), consistent with previous findings.
  • Behavioral: Participants' choices were significantly influenced by the coherence (evidence) of both attributes, and this influence shifted towards the more rewarded attribute in later epochs of the experiment. Response times were faster when evidence strongly supported the chosen response.

Limitations and Future Work

The authors acknowledge that their task, while designed to mimic real-world decisions, is still a simplification. They note that distinguishing between cognitive control demands and value-based evaluation signals in dACC can be challenging, as increased salience of a tempting alternative attribute inherently leads to greater uncertainty. Future work could further investigate this by manipulating the task to make switching attention easier or harder, and by exploring how these neural signals adapt to different types of cognitive demands. They also suggest that further research is needed to fully bridge the understanding of how we maintain focus on a task with how we select specific words on a page, implying a broader application of these findings to complex cognitive processes.

Summary updated

Abstract

Abstract Decision-making is typically studied as a sequential process from the selection of what to attend (e.g., between possible tasks, stimuli, or stimulus attributes) to the selection of which actions to take based on the attended information. However, people often gather information across these levels in parallel. For instance, even as they choose their actions, they may continue to evaluate how much to attend other tasks or dimensions of information within a task. We scanned participants while they made such parallel evaluations, simultaneously weighing how much to attend two dynamic stimulus attributes and which response to give based on the attended information. Regions of prefrontal cortex tracked information about the stimulus attributes in dissociable ways, related to either the predicted reward (ventromedial prefrontal cortex) or the degree to which that attribute was being attended (dorsal anterior cingulate, dACC). Within dACC, adjacent regions tracked uncertainty at different levels of the decision, regarding what to attend versus how to respond. These findings bridge research on perceptual and value-based decision-making, demonstrating that people dynamically integrate information in parallel across different levels of decision making. Naturalistic decisions allow an individual to weigh their options within a particular task (e.g., how best to word the introduction to a paper) while also weighing how much to attend other tasks (e.g., responding to e-mails). These different types of decision-making have a hierarchical but reciprocal relationship: Decisions at higher levels inform the focus of attention at lower levels (e.g., whether to select between citations or email addresses) while, at the same time, information at lower levels (e.g., the salience of an incoming email) informs decisions regarding which task to attend. Critically, recent studies suggest that decisions across these levels may occur in parallel, continuously informed by information that is integrated from the environment and from one’s internal milieu 1,2 . Research on cognitive control and perceptual decision-making has examined how responses are selected when attentional targets are clearly defined (e.g., based on instruction to attend a stimulus dimension), including cases in which responding requires accumulating information regarding a noisy percept (e.g., evidence favoring a left or right response) 3-7 . Separate research on value-based decision-making has examined how individuals select which stimulus dimension(s) to attend in order to maximize their expected rewards 8-11 . However, it remains unclear how the accumulation of evidence to select high-level goals and/or attentional targets interacts with the simultaneous accumulation of evidence to select responses according to those goals (e.g., based on the perceptual properties of the stimuli). Recent work has highlighted the importance of such interactions to understanding task selection 12-15 , multi-attribute decision-making 16-18 , foraging behavior 19-21 , cognitive effort 22,23 , and self-control 24-27 . While these interactions remain poorly understood, previous research has identified candidate neural mechanisms associated with multi-attribute value-based decision-making 11,28,29 and with selecting a response based on noisy information from an instructed attentional target 3–5 . These research areas have implicated the ventromedial prefrontal cortex (vmPFC) in tracking the value of potential targets of attention (e.g., stimulus attributes) 8,11 and the dorsal anterior cingulate cortex (dACC) in tracking an individual’s uncertainty regarding which response to select 30–32 . It has been further proposed that dACC may differentiate between uncertainty at each of these parallel levels of decision-making (e.g., at the level of task goals or strategies vs. specific motor actions), and that these may be separately encoded at different locations along the dACC’s rostrocaudal axis 32,33 . However, neural activity within and across these prefrontal regions has not yet been examined in a setting in which information is weighed at both levels within and across trials. Here we use a value-based perceptual decision-making task to examine how people integrate different dynamic sources of information to decide (a) which perceptual attribute to attend and (b) how to respond based on the evidence for that attribute. Participants performed a task in which they regularly faced a conflict between attending the stimulus attribute that offered the greater reward or the attribute that was more perceptually salient (akin to persevering in writing one’s paper when an enticing email awaits). We demonstrate that dACC and vmPFC track evidence for the two attributes in dissociable ways. Across these regions, vmPFC weighs attribute evidence by the reward it predicts and dACC weighs it by its attentional priority (i.e., the degree to which that attribute drives choice). Within dACC, adjacent regions differentiated between uncertainty at the two levels of the decision, regarding what to attend (rostral dACC) versus how to respond (caudal dACC).
BibTeX
@inproceedings{shenhav2018dissociable,
  author = {Amitai Shenhav and Mark A. Straccia and Jonathan D. Cohen and Matthew Botvinick},
  title = {{Dissociable neural mechanisms track evidence accumulation for selection of attention versus action}},
  booktitle = {Nature Communications},
  year = {2018},
  doi = {10.1038/s41467-018-04841-1},
}
Figure