Contribution of the Frontal Eye Field to Adaptive Target Selection

Abstract

Saccadic target selection is one of the most frequent decisions made by primates and involves multiple brain areas including the frontal eye field (FEF). The FEF also receives a dopaminergic input suggesting an exclusive role of the FEF in value-based and adaptive target selection. This role, however, has been largely unclear because of multiple other brain areas involved in these processes. To address this question, we combined computational modeling and electrical microstimulation of the FEF during a dynamic value-based decision-making task in monkeys. As expected, on stimulated trials, monkeys’ choices were biased toward the target in the response field of the microstimulated site (T_in). On non-stimulated trials following microstimulation, however, monkeys showed a bias away from T_in that slowly decreased as a function of the time since the last stimulation, pointing to the presence of a compensatory mechanism. Moreover, FEF microstimulation and reward history interacted to guide saccadic choice. By simulating choice behavior using multiple network models, we found that our experimental observations can be captured by a model in which microstimulation and value signals interact competitively through a reward-dependent plasticity rule. Together, the results illustrate that reward-dependent modulations of different signals within the FEF could enable this area to contribute to adaptive target selection.

Keywords: reward, learning, microstimulation, oculomotor system