Abstract:
Spatial scale of attention influences saccade latencies.
Madelain, Harwood, Krauzlis & Wallman, VSS 2004
Purpose: We previously showed that, when viewing two concentric rings (0.8 and 8 deg diameter) that together undergo a step displacement (< 8 deg), subjects make saccades at shorter latencies if they attend to the inner ring. We proposed that saccades are triggered faster when the target object leaves the spotlight of attention. We now test three other explanations. First, the difference in latency could simply be an effect of stimulus eccentricity. Second, subjects may divide their attention when instructed to attend to the large ring, causing an increase in latency. Third, ring size may determine latency regardless of whether the stimulus steps outside of the field of attention.
Experiment 1: Large and small stimuli with the same eccentricity. Two pairs of lines, one spanning 2 deg, the other 8 deg, oscillated side-to-side 2 deg from the fixation point. At a random time, they stepped and the lines briefly changed orientation. Subjects reported whether the lines of the assigned pair were transiently parallel. In the three subjects, saccades occurred later when attending large (288 ms) than small (205 ms).
Experiment 2: Rings of various sizes stepping by various distances. At a random time, a single, segmented ring stepped, and briefly changed its number of segments. Subjects reported the number of segments during the change. We varied the size of the ring (from 2 to 8 deg) and the amplitude of the step (from 1 to 9 deg). For each ring size, the saccade latencies were longer for steps much smaller than the ring diameter and shorter for steps as large or larger than the ring diameter.
Conclusions: Neither eccentricity nor divided attention nor ring size can explain the changes in saccade latency we observed. We conclude that saccades are delayed if the position error is smaller than the size of the field of attention.
Poster presented at VSS 2004 (pdf version, 470 K)
Spatial
scale of attention modulates saccade latency
Harwood, Madelain, Krauzlis & Wallman, VSS 2003
Purpose: During smooth
pursuit of a compound stimulus (2 concentric rings, one 0.8 deg, the other
8 deg), the latency of saccades to a 1.5 deg perturbation is shorter when
the spatial scale of attention is narrow (directed to small ring) than
when it is broad (Madelain,
Krauzlis & Wallman, Society for Neuroscience, 2001). A
possible explanation for this attentional effect is that a narrow field
of attention differentially increases the saliency of the small ring,
thereby reducing its spatial uncertainty and leading to corrective saccades
more quickly because the position error is detected more quickly. If this
is so, increasing the contrast of the inner ring should have the same
effect as attending to it.
Methods: On each trial the compound stimulus moved at 10 deg/sec until,
at some point, its path was perturbed forward or backward by 1.5 deg,
and, simultaneously, the number of radial segments in each ring briefly
changed. Three subjects reported the number of segments in either the
small or large ring, according to an auditory cue before each block of
24 trials. We measured the saccade latencies in response to the perturbation
on 1344 trials for each subject.
Results: Catch-up saccades occurred much later when attending to the large
ring than to the small, whether the contrast of the inner ring was 0.2,
0.4 or 0.8 (mean latency: large ring, 437, 427, 436 msec, respectively;
small ring, 175,165, 161 msec).
Conclusions: Our results show that saccade latencies are shortened much
more by attending to the small ring than by increasing its contrast. Therefore
the shorter latencies when attention is narrowed are not due to attentional
modulation of saliency. As an alternative explanation, we propose that
short-latency corrective saccades are generated when the target leaves
the spotlight of attention; with our perturbations this would have occurred
only when the scale of attention was narrow.
Poster presented at VSS 2003 (pdf version, 172 K)
Poster presented at SFN 2001 (pdf version, 5 Mg)
VSS abstracts are published in Journal of Vision 
Sally McFadden, Afsheen Kahn and Josh Wallman published a paper related to these ideas but involving saccadic gain adaptation in Vision Research (Vision Research, 2002, 42, 2709-2726).