The locust's use of motion parallax to measure distance

Publisher Website
Google Scholar
Abstract
Locusts perform a visual behavior known as peering before jumping to targets. Peering consists of a side to side translational movement of the head. Wallace (1959) first proposed that this behavior is performed in order to obtain depth information through motion parallax. The present study analyzes more quantitatively the role of peering in the locust's (Schistocerca americana) estimation of distance in an attempt to understand the neural mechanisms involved. Jump velocity was found to be related monotonically to target distance and was used as a measure of the locust's judgement of target distance. By presenting locusts with a target which moved laterally while they peered (artificial parallax), it was possible to simulate the motion parallax of a target at any distance. Jump velocities elicited by means of artificial parallax were the same as jump velocities elicited by the corresponding real distances, demonstrating that locusts use motion parallax as a cue to distance. By moving the target in the same direction but further than the locust's head it was possible to simulate targets whose motion parallax specified a position behind the animal. Locusts jump forward to such paradoxical stimuli with the jump velocity appropriate for targets located at the absolute value of the simulated distance. This suggests that locusts are insensitive to the relation between the direction of head and image motion. Measurement of jump velocity in locusts with one eye occluded revealed an interaction between the parallax signals from the two eyes — locusts appear to sum (or average) the motion perceived in each eye. Jump velocity is predicted more accurately by target distance than by either image displacement or image velocity during peering. This implies that the locust's computation of target distance involves signals concerning its own head motion.