Behavioural demands direct attentional strategies: the 1993 Banff Annual Seminar in Cognitive Science

Canadian Psychology, Jan 1994 by Shapiro, Kimron, Dixon, Peter

The Banff Annual Seminar in Cognitive Science (BASICS) was founded in 1982 and has met each spring since then in Banff, Alberta, Canada. This past May thus marks the 12th year of this conference where now more than 70 hour - long talks have been presented. BASICS was originated to provide an informal atmosphere for the in - depth discussion of a wide variety of research topics within the broadly defined domain of cognitive science. Topics of discussion have ranged from both animal and human basic perceptual processes throughout the information - processing spectrum to include attention, memory, reading and even the neural substrates of these processes. A significant number of talks over the years has delivered the message that sensory perception and cognition need to be studied together, as each has much to contribute to an understanding of the other. The structure of the BASICS conference provides each speaker with 90 minutes to present his or her research which includes an extensive question - and - answer period. This format has been received enthusiastically by speakers and audience alike and creates a workshop - type atmosphere, rather than one of a typical conference. Although the talks this year ranged in scope, as they have in previous years, visual attention was clearly the dominant topic with speakers addressing the strategic role of attention in spatial and temporal search of briefly presented visual displays, vocabulary acquisition, the estimation of event frequency, and as revealed by the study of specific neuropsychological disorders. The purpose of the present paper is to summarize and find commonalties among the individual abstracts which follow.

The first speaker, Norman Brown (University of Alberta) began his talk with the premise that a valid model of the estimation process is essential to an understanding of the process of encoding and representing event frequency. Furthermore, Brown contended that one has only to look to research conducted on memory to realize that accurate assessment of event frequency depends in large part on the nature of the task and that humans are capable of using a number of strategies, the choice of which will determine their degree of success. To examine this contention, this researcher presented lists of pairs of words comprised of category labels and exemplars and varied whether each category label was paired with the same or a different exemplar. In two experiments which required estimation and reaction time, respectively, subjects had to estimate how frequently each label in the list had appeared. Brown reported that the success of this task was influenced heavily by the strategy chosen and that such strategic choice accounted for similar outcomes previously unexplained by prior researchers.

Visual Attention

In the second talk, John Duncan (MRC - Applied Psychology Unit) continued the theme as to how psychological processes are selected according to the current demands of behaviour. Duncan employs both behavioural and neurophysiological techniques to study visual attention and for years has advocated that such attention is directed at the level of whole objects and that attentional systems coordinate their actions to facilitate object processing. To obtain physiological support for this latter assertion, Duncan and his colleagues recorded from single units in Macaque monkeys and in addition to finding facilitation of cells selectively responsive to the target, also found suppression of responding to nontarget stimuli prior to a saccade to a target in a visual search paradigm. Yet in describing other studies, Duncan revealed that PET measurements from regions of visual cortex involved in object discrimination tasks show that whole object representation is not revealed at this level. Finally, Duncan also reported on collaborations with other colleagues showing that the interference produced by the act of selecting an object for report lasts on the order of 10 times that previously believed to exist for visual displays requiring serial search.

The third talk of the first day of the conference was by Nancy Kanwisher (UCLA). This investigator has been in the forefront of the journals lately for her work with the phenomenon known as repetition blindness (RB). The basic RB outcome is that a repeated target item, such as the second occurrence of a word in a sentence, fails to be reported on a large percentage of trials, even when the failure to attend the second target word seriously affects the grammaticality of the sentence. The studies reported by Kanwisher at BASICS were directed at exploring the relationship between various attributes of the two targets to determine if RB only occurs when there is a match between the attended and the reported attributes. Those in attendance at BASICS learned that this work represented a follow - up of several collaborative efforts investigating this same question with patients suffering from parietal lobe damage. Kanwisher reported to us that she first replicated her basic RB finding using displays containing two simultaneously presented items. In subsequent studies using the same method, she showed that the RB effect occurred only when repetitions of the reported stimulus dimension occurred and that the unreported dimension did not affect performance. In a follow - up study, it was then shown that the first item does not even have to be reported but must be attended. Kanwisher informed us that she is still investigating which aspects of the first target must be attended to produce RB. Thus her talk further supported the theme that the demands of current behaviour are able to be accommodated by the visual attentional system.