Selective Attention for Perception and action:Same or Different

The experiments outlined in this proposal investigate how the visual perceptual demands imposed by a task influence the selection of visual targets for action, and the use of visual information for the guidance of ongoing actions. The relationship between selection-for-perception and selection-for-action is an important issue in visual cognition, which has received much attention in recent years. However, the question has been approached almost exclusively from one direction: the influence of action demands upon perceptual discrimination has been studied, but the reverse relation has not. The experiments will fill this empirical gap, and provide crucial empirical evidence to adjudicate between models of selective attention that assume different degrees of coupling between perception and action (e.g. Milner & Goodale, 1995; Schneider, 1995; Rizzolatti et al, 1994). This issue is not only of theoretical interest, but is also relevant to the design of active user-interfaces in which visual monitoring must occur alongside directed actions. The proposal is for four key experiments, and for the creation of an experimental set-up enabling the future extension of this research programme.

We may attend to a visual object because we want to perceive it more clearly, or because we want to act towards it. A key question in cognitive science is whether perception and action use a shared mechanism of selection. It is known that perception is enhanced for the target of an upcoming hand or eye movement, implying that selection for-perception is yoked to selection-for-action (we tend to look where we are going). The idea of a shared mechanism predicts that objects selected for perception should also be prioritised for action (we should tend to go where we are looking). Our project tested this prediction, using state of the art eye- and hand-movement tracking.

Experiment 1 extended an experiment of Sheliga et al (1995). Participants had to attend, without moving their eyes, to a cue at one side of the screen. The cue told them to which of two targets they should move their eyes. Eye movements deviated away from the cue. Within an established model of movement-coding (Tipper et al, 1997), this repulsion is evidence that an eye movement was automatically prepared to the cue, and actively suppressed It thus implies that selection of the cue for perception entails selection of the same location for action. We also found that the biasing influence of attention was much stronger for vertical than for horizontal eye movements. This special vulnerability of vertical eye movements has been found before for abrupt onset distractors (Van der Stigchel & Theeuwes, 2007; Laidlaw & Kingstone, 2010). Our findings extend the effect to voluntarily attended, non-abrupt distractors.
Experiment 2 transferred this paradigm to a manual reaching task, using eye-tracking to ensure that the eyes remained still. We found qualitatively Similär effects of selection-for-perception on reaching (veering away from the cue), but no imbalance between horizontal and vertical reaches. The special vulnerability of vertical movements is specific to eye movements. Thus, selection-for-perception and selection-for-action are tightly linked, but the expression of this linkage depends upon the effector system used. These conclusions have been presented as a poster at the Experimental Psychology Society (Sandoval Similä, Buonocore & McIntosh. The details of distractor interference differ for eyes and hands. Experimental Psychology Society, London, January 2012), and are being written-up for publication.
Experiments 3 and 4 investigated whether the effect of selection-for-perception on selection-for-action is modulated by the layout of objects. Does it matter whether the cue is within the same object as the movement target? Using vertical eye-movements (as these are most vulnerable to interference), we confirmed that participants are faster to initiate an eye movement when the target is within the same object as the cue, but there was no difference in the spatial biasing of the movement. This suggests a role for object-based selection in reach target selection, but not at the stage of reach programming. These findings are interesting, but further experiments will be required before they are sufficiently well-developed to publish.
Experiments 5 and 6 followed-up on Experiment 2, to study whether selection-for-perception continues to bias selection-for-action throughout a reaching movement. We had participants attend to one or other side of a reaching target, and studied their ability to divert their reach if the target jumped to one side. A strong influence of attention on action was again found, confirming the tight yoking of selection-for-perception and selection-for-action. Selective attention entails movement preparation, and we cannot divert attention away from our movement goals without paying a cost in terms of movement efficiency.