Biomedical Engineering Reference
In-Depth Information
back, early visual processing for both the location and visual features being attended
to.
19.5
Top-down deployment of attention
The precise mechanisms by which voluntary shifts of attention are elicited remain
elusive, although several studies have narrowed down the brain areas primarily in-
volved [8, 20, 23]. Here we focus on two types of experiments that clearly demon-
strate how, first, attention may be deployed on a purely voluntary basis onto one of
several identical stimuli (so that none of the stimuli is more salient than the others),
and, second, how eye movements recorded from observers inspecting a visual scene
with the goal of answering a question about that scene are dramatically influenced
by the question being answered.
19.5.1
Attentional facilitation and cuing
Introspection easily reveals that we are able to voluntarily shift attention towards
any location in our visual field, no matter how inconspicuous that location may be.
More formally, psychophysical experiments may be used to demonstrate top-down
shifts of attention. A typical experiment involves cueing an observer towards one of
several possible identical stimuli presented on a computer screen. The cue indicates
to the observer where to focus on, but only at a high cognitive level (e.g., verbal cue),
so that nothing in the display would directly attract attention bottom-up towards the
desired stimulus. Detection or discrimination of the stimulus at the attended location
are significantly better (e.g., lower reaction time or lower psychophysical thresholds)
than at uncued locations. These experiments hence suggest that voluntarily shifting
attention towards a stimulus improves the perception of that stimulus.
Similarly, experiments involving decision uncertainty demonstrate that if a stimu-
lus is to be discriminated by a specific attribute that is known in advance (e.g., dis-
criminate the spatial frequency of a grating), performance is significantly improved
compared to situations where one randomly chosen of several possible stimulus at-
tributes are to be discriminated (e.g., discriminate the spatial frequency, contrast or
orientation of a grating). Thus, we appear to also be able to voluntarily select not
only where to attend to, but also the specific features of a stimulus to be attended.
These results are closely related to and consistent with the spatial and featural nature
of attentional modulation mentioned in the previous section.
19.5.2
Influence of task
Recording eye movements from human observers while they inspect a visual scene
has revealed a profound influence of task demands on the pattern of eye movements
Search WWH ::
Custom Search