The Neural Basis of Aversive Pavlovian Guidance during Planning

Níall Lally, Quentin J. M. Huys, Neir Eshel, Paul Faulkner, Peter Dayan, Jonathan P. RoiserJournal of Neuroscience 2017 · DOI

Abstract

Real-world decisions often involve complex sequences of choices where initial selections constrain future options. Evaluating every possible path is computationally intractable, leading humans to rely on heuristics. One such shortcut is aversive pruning, a reflexive Pavlovian process where individuals neglect parts of a decision space that reside beyond salient negative outcomes. This study identifies the subgenual cingulate cortex (SGC) as the neural driver of this process, distinct from general planning and valuation networks. Furthermore, individual differences in pruning levels correlate with subclinical anxiety and insular cortex sensitivity to loss, providing insights into how negative Pavlovian influences shape goal-directed evaluations.

Introduction

In theory, optimal decision-making requires a full search of all possible future states. In practice, the exponential growth of decision trees makes this impossible. To manage this complexity, the human brain employs heuristics to "prune" the search space. Aversive pruning is a particularly powerful heuristic: when a path begins with a significant loss, we often stop considering it entirely, even if it eventually leads to a substantial reward. While computationally efficient, this can lead to suboptimal choices. This paper explores the neural mechanisms underlying this reflexive guidance and its relationship to individual psychological traits.

Methods and Task Design

The researchers recruited 40 healthy volunteers to perform a sequential planning task while undergoing functional magnetic resonance imaging (fMRI). Participants navigated a hexagonal maze, making a series of moves to maximize monetary earnings. Each node in the maze was associated with a specific reward or loss, ranging from small gains to large penalties (-70p).

The state-space transition diagram of the hexagonal maze task, showing the possible moves (Left/Right) between nodes.

To isolate the pruning process, the team used computational modeling to compare participant behavior against several reinforcement learning strategies. The winning model included a specific "pruning" parameter ($\gamma_S$), representing the probability of terminating a search upon encountering a loss. This allowed the researchers to use Kullback–Leibler (KL) divergence—a measure of the difference between a full search and a pruned search—as a regressor for the fMRI data.

Computational Mechanisms of Pruning

Aversive pruning acts as a "fast and frugal" shortcut. By truncating branches of the decision tree that start with negative outcomes, the brain reduces the cognitive load required for planning. However, this reflexive shutdown can be "irrational" if the pruned branch contains high-value rewards that outweigh the initial loss.

A decision tree illustrating how aversive pruning truncates the search space at large negative outcomes (e.g., -70), preventing the discovery of later rewards.

Key Results and Neural Dissociation

The study revealed a clear neural dissociation between three fundamental components of sequential decision-making:

  1. General Planning: Recruitment of the cerebellum, dorsolateral prefrontal cortex (DLPFC), and thalamus was associated with the general computational demands of the maze.
  2. Valuation: The ventral striatum and medial orbitofrontal cortex (OFC) tracked the expected monetary value of chosen paths.
  3. Aversive Pruning: The subgenual cingulate cortex (SGC) was robustly and specifically recruited when planning was influenced by aversive pruning. SGC activity increased linearly with the computational "inclination to prune," acting as a bridge between reflexive emotional responses and high-level goal-directed planning.

Behaviorally, participants were significantly less accurate (37% vs. 78%) when the optimal path required passing through a large loss. Pruning-related errors were also associated with faster reaction times, confirming the heuristic's role as a rapid, reflexive shortcut.

Individual Differences and Clinical Implications

Individual variation in pruning behavior was linked to subclinical anxiety levels. Participants with higher anxiety showed a greater tendency to prune and demonstrated stronger insular cortex responses to the actual receipt of monetary losses. This suggests that hypersensitivity to negative outcomes drives the reflexive "shutting down" of future planning.

These findings have significant implications for psychiatry. In conditions like clinical depression or anxiety disorders, the SGC is often found to be hyperactive. This study suggests that such hyperactivity might lead to "over-pruning," where patients prematurely discard potential life paths or solutions because they perceive only immediate negative hurdles, effectively trapping them in suboptimal states.

Conclusion

Aversive pruning is a fundamental Pavlovian mechanism that guides complex goal-directed planning. By identifying the SGC as the key neural signature of this process, this research provides a biological framework for understanding how emotional reflexes shape high-level cognition and how these processes may become maladaptive in psychiatric disorders.

Summary updated

Abstract

Important real-world decisions are often arduous as they frequently involve sequences of choices, with initial selections affecting future options. Evaluating every possible combination of choices is computationally intractable, particularly for longer multistep decisions. Therefore, humans frequently use heuristics to reduce the complexity of decisions. We recently used a goal-directed planning task to demonstrate the profound behavioral influence and ubiquity of one such shortcut, namely aversive pruning, a reflexive Pavlovian process that involves neglecting parts of the decision space residing beyond salient negative outcomes. However, how the brain implements this important decision heuristic and what underlies individual differences have hitherto remained unanswered. Therefore, we administered an adapted version of the same planning task to healthy male and female volunteers undergoing functional magnetic resonance imaging (fMRI) to determine the neural basis of aversive pruning. Through both computational and standard categorical fMRI analyses, we show that when planning was influenced by aversive pruning, the subgenual cingulate cortex was robustly recruited. This neural signature was distinct from those associated with general planning and valuation, two fundamental cognitive components elicited by our task but which are complementary to aversive pruning. Furthermore, we found that individual variation in levels of aversive pruning was associated with the responses of insula and dorsolateral prefrontal cortices to the receipt of large monetary losses, and also with subclinical levels of anxiety. In summary, our data reveal the neural signatures of an important reflexive Pavlovian process that shapes goal-directed evaluations and thereby determines the outcome of high-level sequential cognitive processes.<b>SIGNIFICANCE STATEMENT</b> Multistep decisions are complex because initial choices constrain future options. Evaluating every path for long decision sequences is often impractical; thus, cognitive shortcuts are often essential. One pervasive and powerful heuristic is aversive pruning, in which potential decision-making avenues are curtailed at immediate negative outcomes. We used neuroimaging to examine how humans implement such pruning. We found it to be associated with activity in the subgenual cingulate cortex, with neural signatures that were distinguishable from those covarying with planning and valuation. Individual variations in aversive pruning levels related to subclinical anxiety levels and insular cortex activation. These findings reveal the neural mechanisms by which basic negative Pavlovian influences guide decision-making during planning, with implications for disrupted decision-making in psychiatric disorders.
BibTeX
@article{lally2017neural,
  author = {Níall Lally and Quentin J. M. Huys and Neir Eshel and Paul Faulkner and Peter Dayan and Jonathan P. Roiser},
  title = {{The Neural Basis of Aversive Pavlovian Guidance during Planning}},
  journal = {Journal of Neuroscience},
  year = {2017},
  doi = {10.1523/jneurosci.0085-17.2017},
}
Figure