Abstract
An extensive body of literature shows that flanking elements make it more difficult to perceive target information (e.g. Bouma, 1970; He et al., 1996). We investigated whether this effect, known as crowding, affects the search time in a search task in which eye movements are required to inspect the whole display.
In a one-dimensional search strip, 6 subjects searched for an ‘O’ (gap 0°) amongst 29 ‘Cs’ (gap 0.33°). One-dimensional ‘mask’ strips of Cs were added above and one below the search strip. Subjects were informed that these mask strips never contained the target. In three conditions, we increased the similarity of the masks (and not the search elements) to the target by decreasing the size of the mask gaps (0.33°, 0.17°, 0.03°), thereby increasing crowding (Nazir, 1992; Kooi et al., 1994). A fourth condition did not contain mask strips. Eye movements were measured.
We validated our stimulus psychophysically. We confirmed that increasing target-mask similarity increased the crowding effect: the maximum eccentricity at which the target could be resolved was smaller when target and mask were more similar (i.e. smaller mask gap size).
Turning to the search experiment, with increasing crowding we found: longer search times, more fixations, shorter saccades and longer fixation durations.
From this we conclude that crowding is a bottleneck on saccadic search besides, for example, the distribution of spatial resolution and the scanning rate of attention. The adjustment of the saccade amplitude and the number of fixations is interpreted as due to decreasing visual span size with increasing crowding.