The long-term goal of our research is to understand the neuronal processes that mediate visual perception and visually guided behavior. To this end we are conducting parallel behavioral and physiological experiments in animals that are trained to perform selected perceptual or eye movement tasks. By recording the activity of cortical neurons during performance of such tasks, we gain initial insights into the relationship of neuronal activity to the animal's behavioral capacities. Hypotheses concerning this relationship are tested by modifying neural activity within local cortical circuits to determine whether behavior is effected in a predictable manner. Computer modelling techniques are then used to develop more refined hypotheses concerning the relationship of brain to behavior that are both rigorous and testable. This combination of behavioral, electrophysiological and computational techniques provides a realistic basis for neurophysiological investigation of cognitive functions such as perception, memory and motor planning.

Many of the perceptual tasks employed in the lab make use of a random dot stimulus. This stimulus consists of some small dots moving in a specified direction (signal dots) interspersed with others moving randomly (noise dots). By changing the ratio of signal to noise dots we can change the strength of the motion signal they carry. Perceptually, as the ratio of signal to noise dots (coherence) decreases, it becomes harder to tell in which direction the signal dots are moving. Below are a few Quicktime movies of random dot stimuli at different coherences. The quicktime player is needed to view these movies.