Perceptual visual space is not compressed before saccades
P Reeve,
J Clark, J K O'Regan¶ (Centre for Intelligent Machines, McGill University,
3480 University Street, Montréal, Québec H3A 2A7, Canada;
¶ Laboratoire de Psychologie Expérimentale, Université de Paris V, 71
avenue Edouard Vaillant, F 75014 Boulogne-Billancourt, France)
In recent years, a group of studies has purported to show that, in the brief
period before saccades, perceptual visual space is compressed toward the target
of the saccade. These findings have been speculatively linked to issues of
perceptual stability and to observed visual receptive-field shifts in parietal
cortex and superior colliculus. These studies have focused on errors in
judgments of the location of isolated, brief pre-saccadic flashes or bars,
inferring an instantaneous distortion in the shape of visual space from the
spatial distribution of their mislocalisations. Some have claimed to bridge the
gap between these judgments about isolated points and the overall structure of
visual space by showing that both the width and the number of a set of
pre-saccadically presented objects are underestimated. We show that this results
from the failure to perceive peripheral objects in the set, an event whose
probability increases with stimulus width. We thus show that visual space,
whether judged by the distance between edges of single objects or by distributed
sets of objects, is not compressed before saccades. This demonstrates the need
for a new understanding of how pre-saccadic perceptual mislocalisations and the
structure of visual space are related.
© 2004
Pion Ltd