Parallel representation of context and multiple context-dependent values in ventro-medial prefrontal cortex

bioRxiv (Cold Spring Harbor Laboratory)(2021)

Cited 1|Views3
No score
Abstract
Value representations in ventromedial prefrontal-cortex (vmPFC) are known to guide the choice between options. But the value of an option can be different in different task contexts. Goal-directed behavior therefore requires to know the current context and associated values of options, and to flexibly switch between value representations in a task-dependent manner. We tested whether task-relevant and -irrelevant values influence behavior and asked whether both values are represented together with context signals in vmPFC. Thirty-five participants alternated between tasks in which stimulus color or motion predicted rewards. As expected, neural activity in vmPFC and choices were largely driven by task-relevant values. Yet, behavioral and neural analyses indicate that participants also retrieved the values of irrelevant features, and computed which option would have been best in the alternative context. Investigating the probability distributions over values and contexts encoded in multivariate fMRI signals, we find that vmPFC maintains representations of the current context, i.e. task state, the value associated with it, and the hypothetical value of the alternative task state. Crucially, we show that evidence for irrelevant value signals in vmPFC relates to behavior on multiple levels, competes with expected value signals, and interacts with task state representations. Our results thus suggest that different value representations are represented in parallel and imply a link between neural representations of task states, their associated values and their influence on behavior. This sheds new light on vmPFC’s role in decision making, bridging between a hypothesized role in mapping observations onto the task states of a mental map, and computing value expectations for alternative states.
More
Translated text
Key words
prefrontal cortex,parallel representation,context-dependent,ventro-medial
AI Read Science
Must-Reading Tree
Example
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined