Abstract
Saccade programming can be altered by the appearance of a distractor or a countermanding stimulus. Here we look at the effect of changing target size. Saccade latencies are longer when the target steps within the attentional window, than when stepping outside it. We use this difference to measure the time it takes to change the scale of attention, by having a stimulus change size at different times before or after it makes a 2 deg horizontal step.
The stimulus was a rotating, segmented ring either 1 or 8 deg in diameter. The number of segments briefly changed at a pseudorandom time during the trial, and subjects reported this number (facilitated by having attention at the appropriate spatial scale) and made a saccade to the target step. In half of the trials, the size of the target was changed. We compared the saccadic latencies of the size-change trials to those of the non-size-change trials.
We observed that if the target size-change occurred more than 140 ms before the saccade, the saccade latency reflected the latency appropriate to the new target-size. Inhibition and facilitation of saccades were found, respectively, at times nearer or further from saccade initiation. Interestingly, for large-to-small size-changes we found latencies intermediate between those appropriate for the two target-sizes, consistent with an analog rescaling of the attentional window.
Our data place limits on the temporal dynamics of the putative mechanism of attentional rescaling (zoom model). Alternatively, we present a motor re-planning model: A linear decision signal accumulated at either high (small target) or low (large target) rates, with size-changes modeled either as canceling the existing plan, or continuing at the rate appropriate to the new size. This simple scheme was able to reproduce all the key features of the data.
Meeting abstract presented at VSS 2012