Multiple, interrelated projects are underway in each of three primary lines of research: the neural basis of target selection, the neural basis of control and monitoring of saccadic eye movements, and stochastic models of cognition. Related projects are also being investigated in the form of collaborations with other groups.
Neural Basis of Target Selection
- In monkeys performing search for T (or L) among L’s (or T’s) we are recording from multiple electrodes in FEF and on the surface of the head. We are investigating the timing relationships among neurons and between spikes, local field potentials and event-related potentials.
- We have demonstrated that macaque monkeys have a homologue of the human N2pc.
- In monkeys performing the search-step task we are recording spikes and local field potentials from multiple electrodes in FEF. This task allows us to distinguish the contributions of different classes of neurons and to probe the timecourse of target selection, attention allocation and saccade preparation.
Neural Control and Monitoring of Saccadic Eye Movements
- In monkeys performing a saccade stop signal task we are recording from multiple electrodes in FEF, SEF, ACC and on the surface of the head. Using the spiking, local field potential and event-related potentials, we are investigating several questions with these data:
- How do neurons in FEF control when to initiate saccades? When monkeys adjust their performance of the task by speeding or slowing, how is this accomplished? Our data indicates that this is accomplished through a change in the onset of growth of activity.
- Do neurons in SEF and ACC also control saccade initiation. Our data indicates that they do not. Instead neurons signal errors, reinforcement and conflict.
- Do monkeys have a homologue of the error-related negativity? Our data indicates that they do.
- How do spikes, field potentials and surface potentials relate to one another?
- Future work will use a choice response stopping task and will investigate the influence of specific reward schedules.
Stochastic models of cognition and neurons
- Can the evolving, selective activity of visual neurons in FEF be accumulated to produce saccades? Using as input the activity of the visual neurons during visual search, we are exploring a series of nested models and distinct architectures to gain insight into the sensory-motor mapping process.
- An interactive race model of stop signal task performance was developed and is being refined. For example, how can executive control influence the interactive race model of countermanding? We are exploring how conflict, reward rate and memory retrieval each might contribute to the adjustments of performance expressed by humans and monkeys.
- We showed that performance in the search-step and double-step saccade task is accounted for by a race between the GO processes producing the saccades with a STOP process that interrupts the first GO process. We are now developing an interactive race model for this task.
Collaborations
- A collaboration with Sohee Park and Katy Thakkar at Vanderbilt is investigating the performance of schizophrenics on the saccade stop signal task. The goal is to gain a deeper insight into the disorder through the detailed physiological and computational information obtained from monkeys and human subjects.
- A collaboration with Xiao-Jing Wang, Chung-Chuan Lo, and Martin Paré is investigating a biophysically plausible spiking neuron network model to perform the stop signal task.