Abstract
Saccade adaptation (McLaughlin 1967) generally transfers from one similarly-sized object to another (Deubel 1995). However adaptation is also specific to a particular saccade vector, so that probes placed at locations orthogonal to the axis of adaptation, in particular, will elicit saccades that are only partially adapted (Straube, Fuchs, Usher & Robinson 1997). Thus we cannot infer from available data the extent of adaptation transfer to a probe object which lies partially in and partially out of the adaptation field of the primary target. We tested this by adapting the saccades of 2 S (macaca mulatta) to a small visual target at 20deg eccentricity. We interleaved two types of probe trials (approx. 15% total): 1) a long, thin orthogonally oriented bar centered on the axis of adaptation at the same eccentricity as the primary target, and 2) small squares with the same location as the ends of this bar. Using saccade adaptation we reduced the amplitude of saccades to the primary target; typically the resultant change in amplitude to the end probes along the axis of adaptation was much smaller. We compared populations of bar probe and primary target saccades with the same mean direction, and found that there was very little adaptation to the bar probe over and above the adaptation which transferred to the end probes. In a control task, we adapted an entire range of targets simultaneously which were arrayed orthogonally to the axis of adaptation so as to cover the extent of the bar probe, which was also presented on interleaved trials. In this condition transfer of adaptation from the primary targets to the bar probe was nearly complete. On the basis of this object-specific selective adaptation we conclude that information about the spatial extent of visual objects is available to the monkey oculomotor system.
National Eye Institute, Keck Foundation, Dana Foundation, Fight for Sight foundation.