Emergence of task-dependent representations in working memory circuits.

Cristina Savin,Jochen Triesch

doi:10.3389/fncom.2014.00057

Abstract

A wealth of experimental evidence suggests that working memory circuits preferentially represent information that is behaviorally relevant. Still, we are missing a mechanistic account of how these representations come about. Here we provide a simple explanation for a range of experimental findings, in light of prefrontal circuits adapting to task constraints by reward-dependent learning. In particular, we model a neural network shaped by reward-modulated spike-timing dependent plasticity (r-STDP) and homeostatic plasticity (intrinsic excitability and synaptic scaling). We show that the experimentally-observed neural representations naturally emerge in an initially unstructured circuit as it learns to solve several working memory tasks. These results point to a critical, and previously unappreciated, role for reward-dependent learning in shaping prefrontal cortex activity.

Highlights

Working memory is defined as the temporary storage of stimulusspecific information during a delay period
Can reward-dependent learning alone explain the wide variety of experimental observations on prefrontal cortex (PFC) encoding? To address this question, we studied the effects of reward-dependent learning on the encoding properties of neurons in a working memory circuit
We found that our model is able to capture key aspects of neuronal dynamics during working memory tasks

Summary

Introduction

Working memory is defined as the temporary storage of stimulusspecific information during a delay period. Classic work in monkeys revealed that single neurons in this region exhibit selective persistent activity during the delay period (Miyashita, 1988; Goldman-Rakic, 1990) and its disruption (by electrical stimulation, or due to distracters) leads to a decay in performance (Funahashi et al, 1989). These early observations have been interpreted as the circuit exhibiting attractor dynamics, which enable a subset of the neurons to maintain high firing throughout the delay after a brief stimulus presentation (Amit and Brunel, 1997; Brunel and Wang, 2001). While experiments generally agree on how information is represented in working memory circuits, i.e., using spatio-temporal patterns of neural activity, exactly what information gets encoded is less clear

Methods

Results

Conclusion