Dataset Information

Heuristic and optimal policy computations in the human brain during sequential decision-making.

ABSTRACT: Optimal decisions across extended time horizons require value calculations over multiple probabilistic future states. Humans may circumvent such complex computations by resorting to easy-to-compute heuristics that approximate optimal solutions. To probe the potential interplay between heuristic and optimal computations, we develop a novel sequential decision-making task, framed as virtual foraging in which participants have to avoid virtual starvation. Rewards depend only on final outcomes over five-trial blocks, necessitating planning over five sequential decisions and probabilistic outcomes. Here, we report model comparisons demonstrating that participants primarily rely on the best available heuristic but also use the normatively optimal policy. FMRI signals in medial prefrontal cortex (MPFC) relate to heuristic and optimal policies and associated choice uncertainties. Crucially, reaction times and dorsal MPFC activity scale with discrepancies between heuristic and optimal policies. Thus, sequential decision-making in humans may emerge from integration between heuristic and optimal policies, implemented by controllers in MPFC.

SUBMITTER: Korn CW

PROVIDER: S-EPMC5780427 | biostudies-literature | 2018 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Heuristic and optimal policy computations in the human brain during sequential decision-making.

Korn Christoph W CW Bach Dominik R DR

Nature communications 20180123 1

Optimal decisions across extended time horizons require value calculations over multiple probabilistic future states. Humans may circumvent such complex computations by resorting to easy-to-compute heuristics that approximate optimal solutions. To probe the potential interplay between heuristic and optimal computations, we develop a novel sequential decision-making task, framed as virtual foraging in which participants have to avoid virtual starvation. Rewards depend only on final outcomes over ...[more]

PMID: 29362449

Similar Datasets

Project description:Many real-world situations require navigating decisions for both reward and threat. While there has been significant progress in understanding mechanisms of decision-making and mediating neurocircuitry separately for reward and threat, there is limited understanding of situations where reward and threat contingencies compete to create approach-avoidance conflict (AAC). Here, we leverage computational learning models, independent component analysis (ICA), and multivariate pattern analysis (MVPA) approaches to understand decision-making during a novel task that embeds concurrent reward and threat learning and manipulates congruency between reward and threat probabilities. Computational modeling supported a modified reinforcement learning model where participants integrated reward and threat value into a combined total value according to an individually varying policy parameter, which was highly predictive of decisions to approach reward vs avoid threat during trials where the highest reward option was also the highest threat option (i.e., approach-avoidance conflict). ICA analyses demonstrated unique roles for salience, frontoparietal, medial prefrontal, and inferior frontal networks in differential encoding of reward vs threat prediction error and value signals. The left frontoparietal network uniquely encoded degree of conflict between reward and threat value at the time of choice. MVPA demonstrated that delivery of reward and threat could accurately be decoded within salience and inferior frontal networks, respectively, and that decisions to approach reward vs avoid threat were predicted by the relative degree to which these reward vs threat representations were active at the time of choice. This latter result suggests that navigating AAC decisions involves generating mental representations for possible decision outcomes, and relative activation of these representations may bias subsequent decision-making towards approaching reward or avoiding threat accordingly.

Dataset Information

Heuristic and optimal policy computations in the human brain during sequential decision-making.

Publications

Heuristic and optimal policy computations in the human brain during sequential decision-making.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets