Transfer of graded information through gated receptivity to widely broadcast signals

Abstract

Making accurate decisions requires the brain to maintain evolving representations of the evidence supporting different possible choices. The population of neurons that maintains this evolving representation may change over the course of the decision-making process. Such changes may result from a switch in the set of neurons representing the relevant behavioral output or a switch in the set of neurons receiving task-relevant information. Recent work has shown how intervening actions like eye movements or navigation to a new location can shift the set of neurons that encode subsequent inputs and outputs. Here, we present a computational model of how analog-valued information that supports an evolving decision can be flexibly transferred between populations of neurons without changing the synaptic connectivity of the underlying network. This is accomplished through a dynamic gating mechanism in which information is widely broadcast throughout a network, but only those neurons receiving a gating signal are receptive to this information. Using this framework, we provide a mechanistic explanation for recent experimental results in which information supporting an evolving decision was shown to be transferred between different populations of parietal cortex neurons during intervening smooth pursuit and saccadic eye movements. This mechanism enables organisms to maintain a continuous decision process despite changing frames of reference, offering a potentially general framework for cognitive continuity in dynamic environments.