The Role of MSTd Neurons in Visually Guided Navigation

Author: Maxwell, Marsha A.

Year: 1998

Degree: Dissertation (Ph.D.)

Advisor: Andersen, Richard A.

Committee Members: Perona, Pietro; Andersen, Richard A.; Allman, John Morgan; Konishi, Masakazu

Option: Biology

DOI: 10.7907/z914-cv41

Abstract

When we move forward in the world, what we see appears to expand. It has been postulated that the focus of this expanding optic flow pattern is used to determine our direction of heading. This is only true if our eyes are still. If we move our eyes to track objects in our environment, this disrupts the retinal image and shifts the focus away from the heading. In order to recover heading, our brains must compensate for this shift. Neurons in primate cortical area MSTd respond to optic flow patterns and a portion shift their selectivity during tracking eye movements in a way that compensates for the retinal focus shift. These neurons appear to be taking into account the effects of eye movements on what we see as we navigate through the world.

When we view an expanding or rotating stimulus, the laminar motion caused by eye movements shifts the center of the retinal motion pattern in the direction of the eye movement (expansions) or perpendicular to the eye movement (rotations). Our early results indicated that expansion- and rotation-selective neurons in area MSTd of the macaque use an extra-retinal signal ( efference copy or proprioception) to compensate for the shift in the retinal motion focus caused by pursuit eye movements. The second half of the thesis compares human abilities to compensate for the effects of eye movements using similar expanding and rotating patterns of random dots. Here we found that humans also made perceptual errors in localizing motion foci. Two psychophysical procedures were used, a stepwise method of adjustment (n=5) and a variant of the QUEST maximum-likelihood estimator (n=3). The extent of compensation varied from 51-182%, (mean 105%) with expanding patterns and 22- 54%, (mean 40%) with rotating patterns. Every observer demonstrated better performance (on average by a factor of 2.6) with expansions than rotations. This finding is interesting in that rotation-selective MSTd cells demonstrate a similar degree of compensation to expansion selective cells (both by ~ 42%). It is possible that the differences between human perceptions of expanding versus rotating patterns may reflect differences in the relative numbers of expansion- and rotation-selective cells.

Files