Action-centred attention in virtual environments.

Abstract

Abstract A recent action-centred model of selective attention holds that attention depends upon the relation between the intended target, distracting stimuli, and the action to be performed (Tipper, Lortie, & Baylis, 1992). In contrast to many earlier studies, where perception and action seem to be dissociated, an action-centred approach stresses that what is perceived depends on how we intend to interact with the environment (Bootsma, 1989). Consequently, selective attention is best studied under conditions analogous to the three-dimensional, real-world action in which humans typically engage (Tipper et al., 1992). Three experiments were conducted to assess the predictions of this model when participants are required to direct action to intended targets located within a computer-generated virtual environment. Taken together, the results suggest that human selective attention is predominately influenced by the degree to which perception and action space is aligned. Specifically, unless a reasonably direct spatial alignment of perception and action is evident, any competing response afforded by irrelevant stimuli is less likely to impede either movement preparation or execution. Until recently, models of human selective attention have largely failed to adequately address a fundamental relation between attention and action. Lately, however, action-centred models of attention have been proposed wherein a clear distinction is drawn between a system of attentional selectivity, limited by an easily exceeded information processing capacity, and a system that is simply confined by physical limitations arising from competing actions to specific responses. In essence, action-based models of attention have defined this selectivity as the processes by which competing responses are inhibited and appropriate actions required to complete a specific goal are selected and accurately executed. Indeed, early proponents of what are now termed action-based models (e.g., Shallice, 1972) have long argued that the two functions of selection and operation (e.g., appropriate effector response to a specific stimulus) are, in fact, controlled by the same system. More recently, supporters of action-centred attentional theories (e.g., Tipper, Lortie, & Baylis, 1992) have suggested that classic information-processing models of attention do not reflect typical human interaction with the environment. They maintain that Gibson (1979) was correct in asserting that the environment in which the organism evolves, and the behavioural requirements necessary to survive in that environment, are critical considerations for any study of attention. Tipper et al. attempted to circumvent the ecological limitations of earlier studies in selective attention by employing tasks that required direct manual contact in a three-dimensional experimental situation. In addition, and perhaps most importantly, the methodology adopted in Tipper et al.'s experiments allowed for examination and comparison of three hypothesized forms of internal representations, proposed as necessary frames of reference for appropriate motor responses directed to stimuli in threedimensional space. As noted, the presence of distractors within a visual array has been shown to produce fairly reliable interference effects that result in generally slower movements and decreased response accuracy. Historically, explanations of these interference effects have centred around some variation of information-processing models, but there remains some disagreement as to the exact locus of this interference. Some have suggested that attention is determined by a twodimensional retinal projection of the stimulus (e.g., Broadbent, 1982; Ericksen & Eriksen, 1974) whereas others (e.g., Downing & Pinker, 1985) contend that the attentional selection process is centred about the viewer and information is selected from three-dimensional space. Recent evidence, however, seems to partially discount this second explanation. …