26th Annual Computational Neuroscience Meeting (CNS*2017): Part 2

We continuously need to coordinate multiple processes to accomplish our current goals. By means of its connection with multiple other areas of the brain, the prefrontal cortex (PF) plays a pivotal role in this cognitive challenge [1]. Neurons in PF are known to represent task-relevant information and to memorize current goals until the proper action can be selected. However, how PF neurons transform the goals into specific actions is not completely understood yet. To address this question, we used an experimental task that resembles a situation in which an object needs to be maintained in memory and then perform a motor plan to reach it. Two monkeys were trained to perform a distance discrimination task [2] while the activity of neurons in PF were recorded. All procedures followed the Guide for the Care and Use of Laboratory Animals (1996, SBN 0-309-05377-3) and were approved by the NIMH Animal Care and Use Committee. The monkeys had to decide which of two stimuli (blue or red) sequentially presented on a screen was farther from a reference point. A working memory period separated the end of the presentation of the second stimulus from the reappearance of both stimuli (goals) that served as a “goal” signal. The positions of the two goals were randomized so that, during the working memory period, the monkeys could not predict the future action to perform. Thus, during the delay period the monkeys had to remember the goal (blue or red) and then to select it by touching the corresponding switch below. We found that only a minority of neurons were involved in both the encoding of the goal in memory and the transformation of it into an action by representing the goal also in this phase. Moreover, in equal proportion, they switched or maintained their goal preference across such transition. Such high probability in the change of preference did not occur in other periods of the task and therefore we interpreted it as a signature of an activity reconfiguration of the PF network due to a transition between different collective states. From a theoretical point of view, the active maintenance of goal information in memory requires some degree of stability whereas, by contrast, the same network needs to be susceptible and flexible enough to adapt to the external changes. To account for such dynamics, we propose that the PF network is composed of bistable cell assemblies with heterogeneous excitability [3] such that both dynamical stability and input susceptibility can be simultaneous expressed. Moreover, we show that, although the neurons that represent the goal both in memory and during the goal to action transformation process are only a minority of all neurons, they can play a fundamental role in the PF activity reconfiguration.