Item

Abstract

The ability to detect motion in our environment is a crucial function of the visual system. Motion processing is usually studied by comparing the activity elicited by various motion stimuli relative to a baseline (“no-movement”) condition. However, during natural vision the sensory input is not broken into a series of discrete presentations that are simply switched on and off. By using a motion adaptation paradigm we studied how stimulation history influences the directional selectivity of single neurons in area V4. We found that V4 neurons which classically would be thought as non-directionally selective can in fact acquire directional selectivity after adaptation. We recorded from area V4 of two monkeys using tetrodes and characterized the directional tuning properties of single units using drifting coherent random dot patterns. In agreement with previous studies we find that the majority of area V4 neurons are weakly tuned to the direction of motion when their properties are characterized using the classical stimulation paradigm. The same neurons though, express stronger directional tuning if previously adapted to a moving stimulus for a period of one second. To quantify the amount of directional information present in the activity of V4 neurons we used a Bayesian population decoding method to predict the direction of motion of the stimulus trial by trial using the activity of a population of neurons in a four hundred millisecond window. The average test error dropped significantly when computed after adaptation. It is important to characterize the properties of neuronal circuits under adaptation to better understand the mechanisms of natural vision.