Abstract
When asked to compare two moving stimuli separated by a temporal delay, observers must identify and remember stimulus direction. We examined the properties of these processes in two macaque monkeys by sequentially presenting two random-dot stimuli, sample and test, in opposite hemifields and introducing a random-motion mask during the delay. By manipulating the spatial location of the mask, its size, temporal parameters, and the time of its presentation during the delay we examined the nature of the remembered stimulus.
We found that the mask was effective only when introduced at the location of the test, about 100–200ms after the start of a 1500ms delay, and when its size was similar to that of the remembered sample. The spatial specificity of the mask effect suggests that the information about the remembered direction presented in one hemifield is transferred to the site of the upcoming test in the opposite hemifield. The effectiveness of the mask early in the delay indicates that this transfer is likely to occur shortly after the offset of the sample. The finding that the properties of the remembered stimulus faithfully reflect signals used to encode it suggests the involvement of motion-processing cortical areas in its storage, even when information has to be transferred to the opposite hemifield. This is consistent with the recent finding that the spatial scale of the mechanism underlying storage of visual motion matches that of MT neurons (Zaksas et al, JNP 2001).