From vision to action: the planning and execution of eye movements

Abstract

As we navigate the visual world we move our eyes about in a ballistic fashion roughly three times every second. This process, known as a saccade, brings the high‐resolution portion of our retina, the fovea, onto an object in the visual world. Circuits involved in the planning and execution of eye movements have been well studied across the brain, especially the brain stem circuits that lead to contraction and relaxation of the extraocular muscles. Relatively less is known about the transition from visual inputs to eye movements, a process that involves the cerebral cortex. In particular, a portion of prefrontal cortex known as the frontal eye fields (FEF) have been implicated in the cortical control of eye movements. This brain region contains visual neurons that respond to a flash of light and motor neurons that respond just prior to an eye movement. This is an ideal locus for studying the visuomotor transformation, because both types of neurons exist in close proximity and because one of the outputs of FEF is a set of descending projections that lead to the movement itself. We hypothesized that populations of visual and motor neurons in FEF were functionally organized in a manner that efficiently transforms visual information into a motor command. We studied this by performing extracellular recordings of populations of FEF neurons using a linear electrode array in alert rhesus macaque monkeys performing a conventional memory guided saccade task in which they made an eye movement to the location where a visual stimulus had previously appeared. This task permits evaluation of the visual response, the motor response, and the delay period in between when the eye movement plan must be generated from the sensory stimulus. We focused on the correlated variability among neurons, because it is a key determinant of how well their activity could be decoded by downstream neurons that drive the eye movement itself. When we analyzed the connections between visual and motor neurons, we found a distinct pattern of results. The overall level of correlated variability between these groups was lower than visual‐visual or motor‐motor pairs, and it changed dramatically depending on the direction of the planned eye movement. A decoding analysis revealed that these changes with eye movement direction were organized in a fashion that would be advantageous to downstream neurons in executing the eye movement. These findings suggest that visual and motor populations of neurons in FEF play a unique role in transforming visual information to motor output.Support or Funding InformationNIH Grants EY022928 and EY008098