Are fixational eye movements adaptive? Two tests of the interaction between photoreceptor sampling, eye movements and psychophysical performance

Authors:

Abstract
Under conditions in which visual performance is limited by sampling, small eye movements that influence the way the optic array is sampled have the potential to change visual behaviours. We use high-resolution, adaptive-optics-enabled (AO) in vivo retinal imaging to measure fixational eye movements during psychophysical tasks, and we analyse responses in relation to the way the retinal stimulus is sampled by the photoreceptor mosaic. In the first experiment, fixation stability is measured as a function of edge blur. Under instruction to maintain fixation on an anisotropic target, participants make fewer eye movements in the direction that maximises temporal flux in retinal stimulation. However, edge blur has little effect on stability, suggesting that the error signal for eye position is based on the centre of gravity of a stimulus, rather than local flux at a boundary. In the second experiment, eye movements are tracked whilst assessing psychophysical thresholds on a sampling-limited spatio-chromatic task. Amongst observers who pass standard colour vision tests, the average ratio of L:M cones is 2:1; however, it can vary substantially between individuals. The data are analysed to test whether eye movements enhance spatial thresholds for L- and M-isolating Landolt Cs and whether this advantage follows sampling-based predictions as a function of L:M ratio. AO retinal imaging enables the study of spatio-temporal sampling of visual stimuli in a way that has not previously been possible. We discuss the effect of fixational eye movements on visual performance and the intriguing possibility that fixational eye movements are adaptive to one's specific retinal cone mosaic.