Modelling Emergent Attentional Properties
We recently introduced a computational model, SAIM (Selective Attention Identification Model), which is capable of simulating visual disorders in brain lesioned patients, including visual neglect and extinction . Here, we report that the same model can both simulate known attentional effects in normal subjects and make novel verifiable predictions. SAIM aims to achieve a translation-invariant object recognition by mapping inputs from their location on the retina to a translation-invariant ”focus of attention”. Inputs are competitively identified by matching to stored templates. When there are multiple items in the field, there is also competition between the items to win the mapping process. With these mechanisms, SAIM can reproduce qualitatively the results of (1) the Eriksen ”flanker” experiment, where RTs increase when a target is flankered by distractors of the opposite response category; and (2) the Posner spatial cueing paradigm, where RTs increase, when the locations of cues do not match the locations of targets. In the cueing paradigm SAIM also predicts that on invalid trails the target is perceived as being shifted more into the periphery (overshoot effect). We have confirmed this prediction experimentally. In SAIM, attentional effects are emergent properties of the competition for limited resources which is needed to achieve a translation invariant object recognition. In humans, there may be no need to posit an explicit attentional system to account for emergent ”attentional” effects on behaviour.
KeywordsRetina Posit Dition
Unable to display preview. Download preview PDF.
- R. Desimone, M. Wessinger, L. Thomas, and W. Schneider. Attentional Control of Visual Perception: Cortical and Subcortical Mechanisms. In Cold Spring Habor Symposia on Quantitative Biology, pages 963–971. Cold Spring Habor Laboratory Press, 1990.Google Scholar
- D. Heinke and G. W. Humphreys. SAIM: A Model of Visual Attention and Neglect. In Proc. of the ICANN’97, Lausanne, Switzerland, pages 913–918. Springer Verlag, 199Google Scholar
- G.W. Humphreys and D. Heinke. Spatial representation and selection in the brain: Neuropsychological and computational constraints. Visual Cognition, 5(1/2), 1998.Google Scholar
- M.C. Mozer. The perception of multiple objects: a connectionist approach. The MIT Press, 1991Google Scholar
- B. A. Olshausen, C. H. Anderson, and D. C. v. Essen. A Neurobiological Model of Visual Attention and Invariant Pattern Recognition Based on Dynamic Routing of Information. J. of Neuroscience, 13(11):4700–4719, 1993.Google Scholar
- B. A. Olshausen, C. H. Anderson, and D. C. v. Essen. A Neurobiological Model of Visual Attention and Invariant Pattern Recognition Based on Dynamic Routing of Information. J. of Computational Neuroscience, 2:45–62, 1995.Google Scholar
- H. Pashler, editor. Attention. Psychology Press, 1998Google Scholar
- M.L Posner, C.R.R. Snyder, and B.J. Davidson. Attention and the detection of signals. J. Exp. Psychol., 109:160–174, 1980Google Scholar