Jun-ichiro Kawahara

The preattentive emperor has no clothes : a dynamic redressing

Preattentive models of early vision have not been supported by the evidence. Instead, an input filtering system, which is dynamically reconfigured so as to optimize performance on the task at hand, is proposed. As a case in point, the authors examined Sagi and Juleszs (1985a) claim that detection tasks are processed preattentively and efficiently (shallow search slopes), whereas discrimination tasks require focal attention and yield inefficient steep slopes. In 5 visual search experiments, efficiency was found to depend not on the nature of the task but on whether the task is single or dual. The second component of a dual task, whether detection or discrimination, is performed inefficiently if it does not fit the configuration of the input system, which had been set optimally for the first component. But, even the second component is processed efficiently if there is enough time to reconfigure the system after processing the first component.

The attentional blink is governed by a temporary loss of control

Identification of the second of two brief targets is impaired at intertarget lags of less than about 500 msec. We compared two accounts of thisattentional blink (AB) by manipulating the number of digit distractors—and hence the lag—inserted among three letter targets in a rapid serial visual presentation stream of digit distractors. On the resource-depletion hypothesis, longer lags provide more time for processing the leading target, thus releasing resources for the trailing target. On the temporary-loss-of-control (TLC) hypothesis, intervening distractors disrupt the current attentional set, producing a trailing-target deficit. Identification accuracy for trailing targets was unimpaired not only at lag 1 (conventional lag 1 sparing) but also at later lags, if preceded by another target. The results supported the TLC hypothesis but not the resource-depletion hypothesis. We conclude that the AB is caused by a disruption in attentional set when a distractor is presented while the central executive is busy processing a leading target.

Attentional requirements in visual detection and identification: Evidence from the attentional blink.

Perception of the 2nd of 2 targets (T1 and T2) is impaired if the lag between them is short (0-500 ms). The authors used this attentional blink (AB) to index attentional requirements in detection and identification tasks, with or without backward masking of T2, in 2 stimulus domains (line orientation, coherent motion). With masking, the AB occurred because T2 was masked during the attentional dwell time created by T1 processing (Experiments 1, 2, and 3). Without masking, an AB occurred only in identification because during the attentional dwell time, T2 decayed to a level that could support simple detection but not complex identification. However, an AB occurred also in detection if T2 was sufficiently degraded (Experiment 4). The authors drew 2 major conclusions: (a) Attention is required in both identification and detection, and (b) 2 factors contribute to the AB, masking of T2 while attention is focused on T1 and decay of the T2 trace while unattended.

Task switching mediates the attentional blink even without backward masking.

When two targets are presented in rapid succession, perception of the second target is impaired at short intertarget lags (100–700 msec). Thisattentional blink (AB) is thought to occur only when the second target is backward masked. To the contrary, we show that task switching between the targets can produce an AB even without masking (Experiments 1 and 3). Further, we show that task switching produces an AB only when the second target does not belong to a class of overlearned stimuli such as letters or digits (Experiments 1 and 4). When the second target is masked, however, an AB is invariably obtained regardless of switching or overlearning. We propose that task switching involves a timeconsuming process of reconfiguration of the visual system, during which the representation of the second target decays beyond recognition, resulting in an AB deficit. We suggest that overlearned stimuli are encoded in a form that, while maskable, decays relatively slowly, thus outlasting the delay due to reconfiguration and avoiding the AB deficit.

The subjective size of visual stimuli affects the perceived duration of their presentation

National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan The perception of time spent looking at a stimulus is lengthened or shortened when its physical attributes, such as area, differ from those of a comparison stimulus. We measured the perceived presentation duration of a visual object whose apparent area was altered by the Ebbinghaus illusion while its physical size remained invariant, so that a central circle surrounded by larger inducers appeared smaller than a same-size central circle surrounded by smaller inducers. The results showed that the perceived duration of presentation for apparently larger circles was longer than that of apparently smaller circles, although the actual area remained invariant across all circles. We concluded that the time perception process receives input from later visual processing.

Long-term Abstract Learning of Attentional Set

How does past experience influence visual search strategy (i.e., attentional set)? Recent reports have shown that, when given the option to use 1 of 2 attentional sets, observers persist with the set previously required in a training phase. Here, 2 related questions are addressed. First, does the training effect result only from perseveration with the currently active set or from long-term learning? Experiment 1 supported the latter alternative: When training and test were separated by up to 1 week, to prevent perseveration across the 2 sessions, the training effect was still obtained. Second, is the learning feature-specific (tuned to a precise set of colors) or more abstract? Experiments 2 and 3 supported the latter: When stimulus colors were switched between training and test to remove the possibility of feature-specific learning, the training effect again was obtained. These experiments indicate that attentional set is largely guided by long-term abstract learning.

Intertrial Inhibition of Focused Attention in Pop-Out Search

When a visual search for a color oddball is performed, responses to target-present trials are affected by the color of items in immediately preceding target-absent trials, a phenomenon known as the distractor-previewing effect (DPE). Specifically, the color of the items in the target-absent trial suppresses responses to a target of that color in the subsequent trial, even though participants report a target feature uncorrelated with color. We believe that this suppression reflects a transient inhibitory effect on focused attention that biases attention away from items that are of the same color as the items in the target-absent trial. Experiments 1–3 show that the DPE is present only in tasks that require focused attention. Experiments 4A and 4B show that the DPE persists even when target-absent displays are masked. Last, Experiment 5 shows that the DPE emerges as early as within the first 100 msec of a target-present trial and is fully in place by the 250-msec mark.

Ignorance is bliss: The role of observer expectation in dynamic spatial tuning of the attentional focus

When two sequential targets (T1, T2) are inserted in an RSVP stream of distractors, perception of T2 is impaired at intertarget lags shorter than 700 msec. Paradoxically, this deficit disappears when T2 is presented directly after T1 (lag-1 sparing). Visser, Bischof, and Di Lollo (1999) found that lag-1 sparing occurs only when T1 and T2 are presented in the same stream. In contrast, Shih (2000) obtained lag-1 sparing with targets in separate streams. Four experiments addressed this inconsistency and revealed lag-1 sparing with targets in different streams, but only when observers had no foreknowledge of T1’s location. We hypothesized that when T1 location is known, attention is focused narrowly on that stream; if T2 then appears in the other stream it is missed, and lag-1 sparing does not occur. When T1 location is not known, attention is focused broadly, encompassing both streams, and lag-1 sparing ensues.

Two noncontiguous locations can be attended concurrently: Evidence from the attentional blink

When two targets (T1 and T2) are inserted in a rapid stream of visual distractors (RSVP), detection/ identification accuracy of T2 is impaired at intertarget lags shorter than about 500 msec. This phenomenon, the attentional blink (AB), has been regarded as a hallmark of the inability of the visual system to process multiple items. Yet, paradoxically, the AB is much reduced when T2 is presented directly after T1 (known aslag-1 sparing). Because lag-1 sparing is said to depend on observers’ spatial attention being set to process the first target, we predicted that if observers are set to monitor two RSVP streams, they could process more than two items; that is, two instances of lag-1 sparing would be obtained concurrently. The results of three experiments indicated that this was the case. When observers searched for two targets in each of two synchronized RSVP streams, lag-1 sparing occurred concurrently in both streams. These results suggest that the visual system can handle up to four items at one moment under RSVP circumstances.

The perceptual and cognitive distractor-previewing effect

The time it takes to respond to an odd-colored target (e.g., a red diamond among green diamonds) is reduced when distractor-colored items in an appropriate geometric configuration (e.g., multiple red diamonds) are previewed in a preceding trial. B. A. Goolsby and S. Suzuki (2002) suggested that this phenomenon, the distractor-previewing effect, occurs because target saliency is increased by global adaptation to the previewed distractors. The present study tested and extended this idea with visual search experiments using color, face, motion, and word stimuli. We found that the distractor-previewing effect can be obtained with all of these stimuli. In particular, we found that the distractor-previewing effect was elicited by prior activation of distractors by word labels, suggesting a high-level locus for the effect.