Attention based visual processes

Canadian Psychology, Feb 1996 by Cavanagh, Patrick

Attention is often compared to a spotlight or focus of enhanced processing. We have found that this gain comes with certain unexpected costs. Specifically, judgements of positions of stimuli surrounding an attended event are distorted, repelled away from the attentional focus. But there is more to attention than just selection. Our work shows that it is better characterized as an image which keeps track of attended objects. We find that the resolution or grain of attention -- the finest packing of targets which still allows an individual item to be identified and tracked -- is much coarser than the smallest visual feature which can be resolved.

Moreover, the mere act of tracking an object with attention appears to create impressions of motion. In this sense, attentive tracking is like a pursuit system that involves attention rather than eye movements. Attention - based motion has a limited capacity (Pylyshyn & Storm, 1988) and does not require low - level motion signals in order to keep track of targets. For example, observers can accurately follow equiluminous colour targets with attention even when low - level mechanisms grossly underestimate their apparent velocity. Observations on superimposed luminance and color gratings (Cavanagh, 1992) show that attention - based motion operates independently of low - level motion. Wertheimer (1912) was the first to propose an attention - based motion process (for apparent motion) when he considered the possibility "that the passing across of attention is the phenomenon of motion" (p. 1080). Wertheimer (1912) also reported attentive tracking of ambiguous, counter - phasing, radial spokes which could be seen to rotate in either direction, depending on the "set and posture of attention" (p. 1069). This early work has been replicated recently (Ramachandran & Anstis, 1983; Culham & Cavanagh, 1994a).

One might argue that an attention - based, feature tracking process is not a motion process. In the extreme example of the minute hand of a watch, "motion" is inferred, not sensed. Indeed, attentive tracking may be inherently position based. However, our evidence shows that tracking supports a true motion process. It exhibits effects of fatigue with properties unlike those for low - level mechanisms, and it demonstrates predictive behaviour that goes well beyond simple position analysis. We (Culham & Cavanagh, 1994b) were able to show that after extended tracking of a counterphase grating, a flickering test appeared to move in the direction opposite to the tracked motion. Unlike aftereffects based on low - level motion, these tracking aftereffects were independent of adapting contrast, and were not confined to the adapted area.

The obvious parallels between attentive pursuit and smooth pursuit eye movements lead to a separation of attention processes into those involving discrete displacements of attention (saccades) and those involved in continuous tracking of targets (pursuit). The most important difference between attentive tracking and smooth pursuit is that attention can follow several targets at once and provide an awareness of their configuration (the image aspect of attention). Our recent experiments show that when the set of targets corresponds a known object with some constraints (such as a Johansson point - light walker), attention may call on animation agents to model these more efficiently. These agents, or sprites, enable tracking of larger numbers of image points than would be possible for unconstrained or unfamiliar groupings.

Cavanagh, P. (1992). Attention - based motion perception. Science, 257, 1563 - 1565.

Culham, J.C. & Cavanagh, P. (1994a). Motion capture of luminance stimuli by equiluminous color gratings and by attentive tracking. Vision Research., 34, 2701 - 2706.

Culham, J., & Cavanagh, P. (1994b). Attentive tracking of a counterphase grating produces a motion aftereffect. Investigative Ophthalmology and Visual Science, 35, 1622.

Pylyshyn, Z.W. and Storm, R.W. (1988). Tracking multiple independent targets: evidence for a parallel tracking mechanism. Spatial Vision, 3, 151 - 224.

Ramachandran, V.S. & Anstis, S.M. (1983). Perceptual organization in moving patterns. Nature, 304, 529 - 531.

Wertheimer, M. (1912) Experimentelle Studien uber das Sehen von Bewegung. Zeitschrift fur Psychologie, 61, 161 - 265. Partial English translation in Thorne Shipley (ed.) Classics in Psychology, 1961, New York: Philosophical Library, pp 1032 - 1089.