Item

Abstract

The transmission of visual signals from eye to brain involves considerable delays in conduction and processing1. Because stimuli of varying intensity or colour can cause different delays, it could be difficult to synchronize events from different parts of a visual scene — in particular, our perception of moving stimuli would consistently trail behind their real locations. But the visual system can circumvent such delays by anticipating the path of moving stimuli. Such motion anticipation was assumed to be controlled by high-level motion areas of the visual cortex. Now, very much to our surprise, Berry et al.2 (page 334 of this issue) report that motion anticipation is already accomplished to a large extent in the retina, by neural circuits that were discovered long ago.
Judging the location of moving objects is important for evading obstacles or predators and for catching prey. These tasks would be almost impossible if the relevant information was delayed. If, for example, we assume a processing delay1 of about 100 ms, an animal (or a car nowadays) moving at a speed of 40 km per hour would be seen more than one metre behind its actual position. To overcome this potentially lethal problem, evolution has selected mechanisms that anticipate the path of motion.
The existence of mechanisms that compensate for visual delays was uncovered by clever psychophysical experiments, which compared the perceived locations of flashed and moving objects3, 6 (Fig. 1). If presented at the same position, flashed objects are seen to trail moving objects by as much as 80 ms. The common interpretation of this phenomenon was that both types of stimuli — flashed and moving — go through the same delays in the eye, and that the position of the predictable, moving stimulus is then corrected by movement-selective mechanisms in areas of the brain concerned with analysing visual motion.