A recurrent neural network model of prefrontal brain activity during a working memory task

Abstract

When multiple items are held in short-term memory, cues that retrospectively prioritise one item over another (retro-cues) can facilitate subsequent recall. However, the neural and computational underpinnings of this effect are poorly understood. One recent study recorded neural signals in the macaque lateral prefrontal cortex (LPFC) during a retro-cueing task, contrasting delay-period activity before (pre-cue) and after (post-cue) retrocue onset. They reported that in the pre-cue delay, the individual stimuli were maintained in independent subspaces of neural population activity, whereas in the post-cue delay, the prioritised items were rotated into a common subspace, potentially allowing a common readout mechanism. To understand how such representational transitions can be learnt through error minimisation, we trained recurrent neural networks (RNNs) with supervision to perform an equivalent cued-recall task. RNNs were presented with two inputs denoting conjunctive colour-location stimuli, followed by a pre-cue memory delay, a location retrocue, and a post-cue delay. We found that the orthogonal-to-parallel geometry transformation observed in the macaque LPFC emerged naturally in RNNs trained to perform the task. Interestingly, the parallel geometry only developed when the cued information was required to be maintained in short-term memory for several cycles before readout, suggesting that it might confer robustness during maintenance. We extend these findings by analysing the learning dynamics and connectivity patterns of the RNNs, as well as the behaviour of models trained with probabilistic cues, allowing us to make predictions for future studies. Overall, our findings are consistent with recent theoretical accounts which propose that retrocues transform the prioritised memory items into a prospective, action-oriented format. Author SummaryMany real-world scenarios require us to manipulate the contents of memory to guide behaviour. For example, when grocery shopping, initially we might keep all the items from our shopping list in mind (i.e., in our short-term memory). However, once we spot the dairy aisle, we might want to extract, or prioritise, the items from our list that belong to this category. The question of how such prioritisation of memory items is achieved in the brain is a topic of active research. A recent study in monkeys provided evidence that initially, individual memory items are kept from interfering with one another by being encoded by different brain activity patterns. However, following a contextual cue (akin to the dairy aisle sign), the short-term memory representations of the prioritised items are reconfigured into a format that collapses across the irrelevant differences and highlights aspects relevant for action. In this study, we modelled the emergence of these representational changes in an artificial neural network model. Our results help explain how such processes can be learnt from experience and why they might emerge in the biological brain.