Microcircuitary of performance monitoring: Laminar structure of visual and conflict monitoring in supplementary eye field

Society for Neuronscience 2018
San Diego, CA, United States

We are continuing our investigation of the microcircuitry of supplementary eye field (SEF), an agranular area supporting encoding of visual stimuli and response conflict. Previous work has described neurons encoding visual cues, reward prediction, and response conflict (Stuphorn et al. 2000). However, the laminar distribution of these signals is unknown. With linear electrode arrays, we sampled neural spiking across all layers of SEF while recording overlying EEG in two monkeys performing the saccade countermanding (stop-signal) task. In this task, monkeys earned fluid reward a constant interval after a secondary tone reinforcement for shifting gaze to a peripheral visual target unless a fixation stop signal appeared. The location of the target cued that either a large or small magnitude of reward could be obtained on the current trial. The assignment of reward magnitude alternated across blocks of ~20 trials. Systematic variation of response time demonstrated monkeys’ sensitivity to the reward value. On ~50% of stop-signal trials monkeys shifted gaze in spite of the stop signal; these were followed by a distinct tone reinforcement. The probability of non-canceled errors increased with stop signal delay, and the response times (RT) of errors were consistently less than the RT on trials with no stop signal. Thus, stop signal reaction time (SSRT) could be determined from the Logan model of a race between GO and STOP processes. Response conflict occurs when the GO and STOP processes are co-active during successfully canceled stop signal trials. We isolated 293 neurons across all SEF layers. Neurons responding to the visual target were concentrated in L2, L3, and upper L5 but absent in L6. Responses were commonly more vigorous to targets associated with smaller reward. Responses also varied with RT. Neurons modulating after SSRT on canceled trials were concentrated in L2, L3, and L5. Overall, neurons with wide spikes were distributed across layers, but those with narrow spikes were concentrated superficially. These findings offer new details and insights to inform the first draft of a cortical microcircuit that enables agranular cortex to exert proactive inhibition in response to the conditions and consequences of performance.