Abstract
Saccade latencies are often found to rise for targets near the fovea. While such findings are typically assumed to be contingent on the saccade amplitude, recent evidence shows that the size of the saccade target also plays an important role (Harwood et al., 2008). That is, latencies for proximal targets increase drastically as a function of the target's size. Here we investigate what underlies this interesting size-latency phenomenon. As the increase in latency stems from the manipulation of a visual property, one potential explanation for this phenomenon is that the decision of where to saccade to next is based on a longer period of accumulating visual input (the visual integration window). Using a luminance selection task in which target-distractor difference varied throughout the trial (Experiment 1) we compared the visual integration window for both smaller and larger targets. Despite considerably longer latencies for larger targets (exceeding 70ms longer for some observers), visual integration windows were similar across target sizes. Therefore, next we asked whether the final stages of saccade preparation take longer. In Experiment 2 we focused on the so-called saccade dead time, the period of the latency during which new input no longer affects the saccade destination. Using a double-step task, we evaluated the observer's ability to update the initial saccade destination when the target steps shortly before saccade execution. Similar dead times were found for both small and large targets. In summary, the phenomenon can best be viewed as increasing the period between the visual integration window and the final stages of saccade preparation. The lengthy delay between the integration of visual input and saccade initiation has been an enduring conundrum. As the current results demonstrate that this delay can be modulated using the size-latency phenomenon, understanding this phenomenon can provide crucial insights into the sensorimotor decision process.
Meeting abstract presented at VSS 2014