Binocular Integration for Behavior in Mice

Abstract

Despite the presence of disparity selective neurons in mouse primary visual cortex, there has been no evidence that mice integrate binocular cues to drive behavior. Previous work has demonstrated that rodents can estimate the size of a gap (Legg & Lambert, 1990; Kerr et al., 2013), but it is unclear from these studies whether the cues that are used are monocular or binocular. To determine whether binocular integration can drive behavior, we trained animals to distinguish between toward and away motion, presented dichoptically, for which the cues were only accessible by integrating information between the two eyes. Mice were trained to walk on a ball suspended in the air, while vertical gratings with spatial frequency of 0.02cpd was presented to each eye. Gratings moving in opposite directions for the two eyes generate motion through depth: "toward" stimuli are generated by presenting the left eye with a leftward grating and the right eye with a rightward grating; "away" stimuli are generated by presenting the left eye with a rightward grating and the right eye with a leftward grating. Mice were trained to walk on "away" stimuli and stop on "toward" stimuli and also walk on "right" stimuli, which was generated by presenting same rightward grating in both eyes. After training, mice are able to distinguish between toward, away, and right motion. Even though the animal was able to accomplish 100% correct on the task for some days, this was very rare. Most of the time the animal was able to get 70-90% correct which varied from day to day. This errors might be due to changes in eye positions or the animal losing focus during the train (such as grooming). This behavioral evidence demonstrates that mice can use binocular cues to drive their behavior. Meeting abstract presented at VSS 2016