Abstract
Previous research has demonstrated a role of the supplementary eye field (SEF) in detecting errors, registering success, and exerting pro-active control on saccade production. The cortical circuitry accomplishing these computations is unknown. We present neurophysiological data from two monkeys collected using a linear electrode array during a visually-guided saccade countermanding task. Monkeys were rewarded for making a saccade to a visual target unless, in infrequent random trials, a stop signal appeared which instructed the subject to cancel this pre-planned saccade. Analysis of performance using a race model provides the duration of a covert STOP process. From 16 perpendicular penetrations, we isolated 293 neurons across all layers. Here, we report the laminar organization of three types of neurons that were modulated most prominently after the STOP process terminated to cancel the planned saccade. Two populations of neurons generated complementary reduction and elevation of discharge rates after the STOP process, resembling movement and fixation neurons found in ocular motor structures. However, because this modulation arose too late to exert reactive control over saccade initiation, we interpret the neurons as enabling and disabling task goals. Enable neurons were dense in L2/3 and upper L5 and commonly had broad spikes. Disable neurons were restricted to L2/3, commonly had narrow spikes, and modulated before Enable neurons. A third population of neurons exhibited pronounced, transient modulation after the STOP process - that scaled with the magnitude of conflict inferred between competing gaze-holding and gaze-shifting neuron populations. Conflict neurons were found in L2/3 and L6 and were a mixture of broad and narrow spikes. The conflict signal arose earliest in L2/3 and later in L6. These findings further delineate the mechanisms of medial frontal cortex in response monitoring, constrain circuit-level models of executive control, and guide inverse modelling solutions of visual performance event-related potentials.
Acknowledgement: This work was supported by R01-MH55806, P30-EY08126, and by Robin and Richard Patton through the E. Bronson Ingram Chair in Neuroscience.