My long range objective is to contribute to the understanding of neuronal correlates underlying cognitive processes related to decision making and flexible response selection in primates. I examine the dynamic and distributed representations of multiple factors, both internal (such as subject preferences and current mental states) and external (task contingencies, reward context), that determine goal-directed behaviors.
The ability to efficiently explore available options and to quickly react to behaviorally relevant changes in the environment is very important for evolutionary and individual success of human and simian primates. Primates are visual creatures - they tend to relay on the visual information to form adequate behavioral motor responses, and eye movement control and visuomotor coordination are crucial for their survival. While we may have a good idea about the basic mechanisms at the sensory input and motor output levels of the neuronal processing, we still lack a clear comprehension of the intervening stages, where the complex interaction between external inputs, current and past state representations take place.
My Ph.D. work focused on a relatively early sensory stage - the neuronal mechanisms of active vision in primary visual cortex (V1) of alert behaving monkeys. Together with my advisors, Max Snodderly and Moshe Gur, I re-examined the basic properties of V1 neurons, such as spatiotemporal linearity and receptive field structure, during active fixation, contrasting it to the previous work done in anaesthetized animals, and found important differences in the manner V1 neurons encode visual stimuli in the alert state. In another project, we studied effects of fixational and voluntary eye movements ("microsaccades", drifts, and larger saccades) on the activity of V1 neurons with and without visual stimulation, revealing two complementary types of visual activation and demonstrating a presence of extraretinal modulation in the primary visual cortex.
My current research centers on four main aspects of higher-order visuomotor functions in the primate brain, employing fMRI in monkeys and humans and neuronal recordings in monkeys. The first is space representation in cortical structures involved in transformation of sensory cues into goal-directed saccades, and the elucidation of similarities and differences between human and monkey encoding of space for action. The second is the dynamics of spatial decision signals and response selection across different brain areas within and across hemispheres, and neural correlates of internal spatial biases that contribute to overt choice behavior. The third is how expected reward is influencing the choice behavior and neural representations of action planning. The fourth is a macaque monkey model of spatial neglect and decision making deficits using combined fMRI and reversible pharmacological inactivation approach.
The first and crucial step in addressing the questions outlined above was the development of time-resolved event-related approach to monkey fMRI. Although now routinely used in human imaging, this technique presents special challenges for the monkey fMRI, especially at the high-field, where static and dynamic imaging artifacts due to magnetic field inhomogeneities and other off-resonance effects introduce strong spatial and temporal distortions into functional data. In particular, even slightest motions of monkey body, limb and jaws, and licking during reward delivery, introduce strong spurious fluctuations into the fMRI time-courses. The use of rigorous training procedures, real-time behavioral control and feedback, incorporation of body and jaw movement controls, and automatic selection of non-contaminated data segments for the event-related analysis allowed us to extract BOLD signal time-courses and to reliably delineate responses from different trial epochs, attributing fMRI activity to specific visual, motor, and cognitive events in the task. This enables us to study the dynamics of neural signals, similarly to a standard electrophysiology approach.