We examined a visual search task, in which observers responded to the high-acuityaspect of a popout target (shape of an odd-colored diamond or vernier offset of an odd spatial-frequency patch). Repetition of the attention-driving feature (color or spatial frequency) in this task primes the popout; repetition of the high-acuity aspect (shape, vernier offset) does not. Priming of pop-out is due to a decaying memory trace of the attention-focusing feature laid down with each trial. The trace exerts a diminishing effect over the following five to eight trials (-30 sec), and its influence over this time is cumulative. Observers cannot willfully overcome the priming, which suggests that it is passive and autonomous. Both target facilitation and distractor inhibition are evident; the former has a greater effect. The phenomenon shows complete binocular transfer.In a typical visual search experiment, the observer's task is to detect the presence of the odd target within a field of distractors. Two well-known effects have been observed. Serial search is the term attributed to the finding that an increasing number ofdistractors increases the time necessary to find the target, yielding positive slope functions (Sagi & Julesz, 1985;Treisman & Gelade, 1980). Parallel search is the term attributed to the finding that the target is detected equally quickly regardless of the number of distractors, yielding flat reaction time versus distractor number functions. In this situation, the target also "pops out"; that is, attention is automatically drawn to the odd item. The flat functions and the presence of"pop-out" were generally assumed to be causally related. Bravo and Nakayama (1992), however, dissociated these two aspects by adding an additional requirement to the task, which tapped into its attention-focusing aspect. They asked observers to respond to the shape of the odd-colored target, not its presence or absence. It was their claim that responding to the shape necessitates the spatial focusing of attention to the odd target. As such, this task examines the characteristics of attention focusing in pop-out. In contrast, the previously used presence versus absence task does not require the focusing of attention and is not informative regarding the deployment of focal attention.Drawing on theoretical accounts of attentional guidance, Bravo and Nakayama (1992) made a number of experimental predictions. To start, they reviewed two general processes that would be useful for focusing atThis work was supported by AFOSR Grant F49620-92-J-0016 to K.N. We thank Shinsuke Shimojo for his comments on an earlier version of this manuscript. V.Maljkovic is now in the Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology. Correspondence should be addressed to K. Nakayama,

No abstract

Human observers fixated the center of a search array and were required to discriminate the color of an odd target if it was present. The array consisted of horizontal or vertical black or white bars. In the simple case, only orientation was necessary to define the odd target, whereas in the conjunctive case, both orientation and color were necessary. A cue located at the critical target position was either visible all the time (sustained cuing) or it appeared at a short variable delay before the array presentation (transient cuing). Sustained visual cuing enhanced perception greatly in the conjunctive, but not in the simple condition. Perception of the odd target in the conjunctive display was improved even further by transient cuing, and peak discrimination performance occurred if the cue preceded the target array by 70-150 msec. Longer delays led to a marked downturn in performance. Control experiments indicated that this transient attentional component was independent of the observers' prior knowledge of target position and was not subject to voluntary control. We provide evidence to suggest that the transient component does not originate at the earliest stages of visual processing, since it could not be extended in duration by flickering the cue, nor did it require a local sensory transient to trigger its onset. Neither the variation in retinal eccentricity nor changing the paradigm to a vernier acuity task altered the basic pattern of results. Our findings indicate the existence of a sustained and a transient component of attention, and we hypothesize that of the two, the transient component is operative at an earlier stage of visual cortical processing.

We tested four people who claimed to have significantly better than ordinary face recognition ability. Exceptional ability was confirmed in each case. On two very different tests of face recognition, all four experimental subjects performed beyond the range of control subject performance. They also scored significantly better than average on a perceptual discrimination test with faces. This effect was larger with upright than inverted faces, and the four subjects showed a larger ‘inversion effect’ than control subjects, who in turn showed a larger inversion effect than developmental prosopagnosics. This indicates an association between face recognition ability and the magnitude of the inversion effect. Overall, these ‘super-recognizers’ are about as good at face recognition and perception as developmental prosopagnosics are bad. Our findings demonstrate the existence of people with exceptionally good face recognition ability, and show that the range of face recognition and face perception ability is wider than previously acknowledged.

Treisman and others have reported that the visual search for a target distinguished along a single stimulus dimension (for example, colour or shape) is conducted in parallel, whereas the search for an item defined by the conjunction of two stimulus dimensions is conducted serially. For a single dimension the target 'pops out' and the search time is independent of the number of irrelevant items in the set. For conjunctions, the search time increases as the set becomes larger. Thus, it seems that the visual system is incapable of conducting a parallel search over two stimulus dimensions simultaneously. Here we extend this conclusion for the conjunction of motion and colour, showing that it requires a serial search. We also report two exceptions: if one of the dimensions in a conjunctive search is stereoscopic disparity, a second dimension of either colour or motion can be searched in parallel.