A signal detection model predicts the effects of set size on visual search accuracy for feature, conjunction, triple conjunction, and disjunction displays

Attentional guidance relies on a winner-take-all mechanism

The finding that attention can encompass several non-contiguous items at once challenges the current models of visual search based on a winner-take-all mechanism assuming the selection of a single object. It has been proposed instead that attentional guidance involves mechanisms selecting all relevant items simultaneously. In order to test this hypothesis, we studied attentional allocation during various visual search tasks. We confirmed that attention can indeed select several items concurrently but on the basis of their spatial relation, not relevance. This finding corroborates the view that during visual search, attentional guidance is based on a winner-take-all mechanism.

The speed of free will

Do voluntary and task-driven shifts of attention have the same time course? In order to measure the time needed to voluntarily shift attention, we devised several novel visual search tasks that elicited multiple sequential attentional shifts. Participants could only respond correctly if they attended to the right place at the right time. In control conditions, search tasks were similar but participants were not required to shift attention in any order. Across five experiments, voluntary shifts of attention required 200-300 ms. Control conditions yielded estimates of 35-100 ms for task-driven shifts. We suggest that the slower speed of voluntary shifts reflects the "clock speed of free will". Wishing to attend to something takes more time than shifting attention in response to sensory input.

Statistical decision theory to relate neurons to behavior in the study of covert visual attention

Scrutiny of the numerous physiology and imaging studies of visual attention reveal that integration of results from neuroscience with the classic theories of visual attention based on behavioral work is not simple. The different subfields have pursued different questions, used distinct experimental paradigms and developed diverse models. The purpose of this review is to use statistical decision theory and computational modeling to relate classic theories of attention in psychological research to neural observables such as mean firing rate or functional imaging BOLD response, tuning functions, Fano factor, neuronal index of detectability and area under the receiver operating characteristic (ROC). We focus on cueing experiments and attempt to distinguish two major leading theories in the study of attention: limited resources model/increased sensitivity vs. selection/differential weighting. We use Bayesian ideal observer (BIO) modeling, in which predictive cues or prior knowledge change the differential weighting (prior) of sensory information to generate predictions of behavioral and neural observables based on Gaussian response variables and Poisson process neural based models. The ideal observer model can be modified to represent a number of classic psychological theories of visual attention by including hypothesized human attentional limited resources in the same way sequential ideal observer analysis has been used to include physiological processing components of human spatial vision (Geisler, W. S. (1989). Sequential ideal-observer analysis of visual discrimination. Psychological Review 96, 267-314.). In particular we compare new biologically plausible implementations of the BIO and variant models with limited resources. We find a close relationship between the behavioral effects of cues predicted by the models developed in the field of human psychophysics and their neuron-based analogs. Critically, we show that cue effects on experimental observables such as mean neural activity, variance, Fano factor and neuronal index of detectability can be consistent with the two major theoretical models of attention depending on whether the neuron is assumed to be computing likelihoods, log-likelihoods or a simple model operating directly on the Poisson variable. Change in neuronal tuning functions can also be consistent with both theories depending on whether the change in tuning is along the dimension being experimentally cued or a different dimension. We show that a neuron's sensitivity appropriately measured using the area under the Receive Operating Characteristic curve can be used to distinguish across both theories and is robust to the many transformations of the decision variable. We provide a summary table with the hope that it might provide some guidance in interpreting past results as well as planning future studies.

The detection of temporally defined objects does not require focused attention

Perceptual grouping is crucial to distinguish objects from their background. Recent studies have shown that observers can detect an object that does not have any unique qualities other than unique temporal properties. A crucial question is whether focused attention is needed for this type of grouping. In two visual search experiments, we show that searching for an object defined by temporal grouping can occur in parallel. These findings suggest that focused attention is not needed for temporal grouping to occur. It is proposed that temporal grouping may occur because the neurons representing the changing object elements adopt firing frequencies that cause the visual system to bind these elements together without the need for focused attention.

Attention and the detection of masked radial frequency patterns: Data and model

A radial frequency (RF) stimulus is strongly masked by a second, surrounding RF stimulus that follows the first after a critical stimulus onset asynchrony (SOA) of around 100ms. We sought to determine whether a mask-dependent attentional cuing effect, like that found when detecting pattern-masked sinusoidal gratings, would be obtained with RF stimuli. Observers detected RF modulations in cued or miscued stimuli that were masked with a simultaneous (SIM) RF mask or a delayed (SUC) RF mask that followed it after 100ms. There were large cuing effects in the SUC condition and small cuing effects in the SIM condition, replicating previous findings. The data are well described by a model in which masks affect the informational persistence of stimuli and cues affect the rate at which stimulus information is transferred into visual short-term memory.

Spatial and temporal dynamics of attentional guidance during inefficient visual search

Spotting a prey or a predator is crucial in the natural environment and relies on the ability to extract quickly pertinent visual information. The experimental counterpart of this behavior is visual search (VS) where subjects have to identify a target amongst several distractors. In difficult VS tasks, it has been found that the reaction time (RT) is influenced by salience factors, such as the target-distractor similarity, and this finding is usually regarded as evidence for a guidance of attention by preattentive mechanisms. However, the use of RT measurements, a parameter which depends on multiple factors, allows only very indirect inferences about the underlying attentional mechanisms. The purpose of the present study was to determine the influence of salience factors on attentional guidance during VS, by measuring directly attentional allocation. We studied attention allocation by using a dual covert VS task in subjects who had 1) to detect a target amongst different items and 2) to report letters briefly flashed inside those items at different delays. As predicted, we showed that parallel processes guide attention towards the most relevant item by virtue of both goal-directed and stimulus-driven factors, and we demonstrated that this attentional selection is a prerequisite for target detection. In addition, we show that when the target is characterized by two features (conjunction VS), the goal-directed effects of both features are initially combined into a unique salience value, but at a later stage, grouping phenomena interact with the salience computation, and lead to the selection of a whole group of items. These results, in line with Guided Search Theory, show that efficient and rapid preattentive processes guide attention towards the most salient item, allowing to reduce the number of attentional shifts needed to find the target.

The role of the angular gyrus in visual conjunction search investigated using signal detection analysis and transcranial magnetic stimulation

Transcranial magnetic stimulation (TMS) delivered over the posterior parietal cortex increases choice reaction times in visual search for a target defined by a conjunction of features. Some recent studies of visual search have taken an approach based on signal detection theory, the findings of which are not addressed by studying the disruptive effects of TMS on reaction time. Here we investigated the role of the posterior parietal cortex in visual search by applying TMS while subjects performed unspeeded feature and conjunction visual search tasks matched for level of difficulty. TMS over the right, but not the left angular gyrus (AG) in the parietal cortex, nor vertex decreased subjects' sensitivity on the conjunction but not the feature search task, as measured by the signal detection measure, d'. Changes in bias, specifically the tendency to make false positive responses, were less clear. We consider the findings in terms of four possible explanation: binding, attentional control, spatial localisation and visuomotor co-ordinate transformations.

Eye movement statistics in humans are consistent with an optimal search strategy

Most models of visual search are based on the intuition that humans choose fixation locations containing features that best match the features of the target. The optimal version of this feature-based strategy is what we term "maximum a posteriori (MAP) search." Alternatively, humans could choose fixations that maximize information gained about the target's location. We term this information-based strategy "ideal search." Here we compare eye movements of human, MAP, and ideal searchers in tasks where known targets are embedded at unknown locations within random backgrounds having the spectral characteristics of natural scenes. We find that both human and ideal searchers preferentially fixate locations in a donut-shaped region around the center of the circular search area, with a high density of fixations at top and bottom, while MAP searchers distribute their fixations more uniformly, with low density at top and bottom. Our results argue for a sophisticated search mechanism that maximizes the information collected across fixations.

Saccades and covert shifts of attention during active visual search: spatial distributions, memory, and items per fixation

Target detection during active visual search was examined. The chance corrected spatial distribution of target detection was found to be symmetrically distributed around the point of fixation and, unexpectedly, was independent of the proximity of fixations to the display boundaries. Memory was found to play a very limited role in target detection, but a significant role in the guidance of eye movements. A model of covert shifts was used to estimate the number and spatial distribution of shifts required to explain observed performance. An increase from one to five shifts per fixation across increasing array set size as estimated by two different methods was inconsistent with unchanging fixation durations, suggesting that multiple covert shifts are not occurring during the fixations in active search.

Evidence against signal enhancement as a mechanism of direct selection by color

Two of the possible mechanisms that have been put forward to explain precuing effects are signal enhancement and uncertainty reduction. Signal enhancement leads to processing advantages for valid information because the signal at the known input channel is enhanced, whereas uncertainty reduction allows observers to ignore confusing distractor items in a display. Both mechanisms have been reported to be involved in location precuing, but it is still unclear which of these two mechanisms is responsible for color cuing effects. Two experiments are reported in which expectancy for a certain color in a single-item display was created. Targetswere presented briefly and were masked. If color cues produce signal enhancement, then a color cuing effect should result. If color cues only allow uncertainty reduction, however, there should be no color cuing effect because there are no distractors in single-item displays. The results of both experiments favor uncertainty reduction as the mechanism behind color cuing, because no signs of signal enhancement-based cuing effects were observed.

Parallel and serial processes in visual search

A long-standing issue in the study of how people acquire visual information centers around the scheduling and deployment of attentional resources: Is the process serial, or is it parallel? A substantial empirical effort has been dedicated to resolving this issue (e.g., J. M. Wolfe, 1998a, 1998b). However, the results remain largely inconclusive because the methodologies that have historically been used cannot make the necessary distinctions (J. Palmer, 1995; J. T. Townsend, 1972, 1974, 1990). In this article, the authors develop a rigorous procedure for deciding the scheduling problem in visual search by making improvements in both search methodology and data interpretation. The search method, originally used by A. H. C. van der Heijden (1975), generalizes the traditional single-target methodology by permitting multiple targets. Reaction times and error rates from 29 representative search studies were analyzed using Monte Carlo simulation. Parallel and serial models of attention were defined by coupling the appropriate sequential sampling algorithms to realistic constraints on decision making. The authors found that although most searches are conducted by a parallel limited-capacity process, there is a distinguishable search class that is serial.

On the costs and benefits of faces and words: process characteristics of feature search in highly meaningful stimuli

The authors present a comprehensive consideration of the process characteristics of visual search in contexts that vary in their meaningfulness. The authors frame hypotheses regarding process architecture, stopping rule, capacity, and channel independence, using analytic results and a rigorously specified dynamic system to characterize a set of alternative hypotheses that vary along all of these dimensions. Results of the tests of these hypotheses suggest that process architecture and the stopping rule do not distinguish the processing of meaningful and meaningless forms. The major distinction between configural and nonconfigural processing was with regard to processing capacity, potentially implicating channel interdependencies. All of these conclusions hold for both faces and words.

An inter-item similarity model unifying feature and conjunction search

We propose a model of visual search to address the hitherto unresolved issue of reconciling serial deployment of attention accounts with inter-item similarity effects. Target-distractor and distractor-distractor similarity were systematically varied in 85 (17x5) set type-size conditions over seven experiments, including univariate feature and bivariate conjunction search. The model, a power (square root) function of dimension-specific target-distractor and distractor-distractor similarity in linear combination with set size, accounted for 98% of the variance on type-size means. It suggests that much of efficient and inefficient search can be unified under a single theory involving item similarity.

A search model and figure of merit for observer data acquired according to the free-response paradigm

Search is a basic activity that is performed routinely in many different tasks. In the context of medical imaging it involves locating lesions in images under conditions of uncertainty regarding the number and locations of lesions that may be present. A search model is presented that applies to situations, as in the free-response paradigm, where on each image the number of normal regions that could be mistaken for lesions is unknown, and the number of observer generated localizations of suspicious regions (marks) is unpredictable. The search model is based on a two-stage model that has been proposed in the literature, according to which, at the first stage (the preattentive stage) the observer uses mainly peripheral vision to identify likely lesion candidates, and at the second stage the observer decides (i.e., cognitively evaluates) whether or not to report the candidates. The search model regards the unpredictable numbers of lesion and non-lesion localizations as random variables and models them via appropriate statistical distributions. The model has three parameters quantifying the lesion signal-to-noise ratio, the observer's expertise at rejecting non-lesion locations, and the observer's expertise at finding lesions. A figure-of-merit quantifying the observer's search performance is described. The search model bears a close resemblance to the initial detection and candidate analysis (IDCA) model that has been recently proposed for analysing computer aided detection (CAD) algorithms. The ability to analytically model and quantify the search process would enable more powerful assessment and optimization of performance in these activities, which could be highly significant.

Spatial and feature-based effects of exogenous cueing on visual motion processing

In two experiments, we investigated the effects of exogenous cueing on visual motion processing. The first experiment shows that the typical pattern of reaction time (RT) effects, namely early facilitation and later inhibition of return (IOR), can be obtained using a color change as exogenous cue and a direction change as target. In the second experiment, we manipulated the validity of the cue independently with respect to location and feature using transparent motion stimuli. Facilitation of RTs with short cue-target interstimulus-intervals (ISIs) was only evident for targets with both the valid location and the valid feature. Furthermore, at longer cue-target intervals, RTs were prolonged for targets at the cued location, irrespective of the cued feature. These results demonstrate spatial and feature-based components of early facilitation and purely spatial IOR.

Attentional selection: A salience-based competition for representation

The role of salience in localized attentional interference (LAI) was examined. In two experiments, target discrimination performance was measured as a function of the spatial separation between the target and a salient distractor item. In Experiment 1, both the salience of the distractor and that of a target were manipulated. Distractor salience was manipulated via size changes to the distractor, and target salience was manipulated by using unmasked or onset targets. When the target was of low salience, the magnitude of interference from the distractor increased with distractor salience. However, when the target had an abrupt onset, the distractor had no impact on target performance. In Experiment 2, the attentional salience of the distractor was manipulated using a probability manipulation. Displays contained both a target and a color singleton distractor. The color singleton produced LAI when it was predictive of the target location but not when it was unpredictive of the target location. The results of both experiments are consistent with models of competition-based attentional selection.

Visual clutter causes high-magnitude errors

Perceptual decisions are often made in cluttered environments, where a target may be confounded with competing “distractor” stimuli. Although many studies and theoretical treatments have highlighted the effect of distractors on performance, it remains unclear how they affect thequality of perceptual decisions. Here we show that perceptual clutter leads not only to an increase in judgment errors, but also to an increase in perceived signal strength and decision confidence on erroneous trials. Observers reported simultaneously the direction and magnitude of the tilt of a target grating presented either alone, or together with vertical distractor stimuli. When presented in isolation, observers perceived isolated targets as only slightly tilted on error trials, and had little confidence in their decision. When the target was embedded in distractors, however, they perceived it to be strongly tilted on error trials, and had high confidence of their (erroneous) decisions. The results are well explained by assuming that the observers' internal representation of stimulus orientation arises from a nonlinear combination of the outputs of independent noise-perturbed front-end detectors. The implication that erroneous perceptual decisions in cluttered environments are made with high confidence has many potential practical consequences, and may be extendable to decision-making in general.

Visual targets imbedded in distractors lead not only to judgment errors, but also to higher confidence about erroneous decisions. The signed max model provides a plausible framework for explaining the disparity between perception and reality.

Attention speeds processing across eccentricity: feature and conjunction searches

We investigated whether the effect of covert attention on information accrual varies with eccentricity (4 degrees vs 9 degrees) and the complexity of the visual search task (feature vs conjunction). We used speed-accuracy tradeoff procedures to derive conjoint measures of the speed of information processing and accuracy in each search task. Information processing was slower with more complex conjunction searches than with simpler feature searches, and overall it was faster at peripheral (9 degrees) than parafoveal (4 degrees) locations in both search types. Covert attention increased discriminability and accelerated information accrual at both eccentricities, and the magnitude of this attentional effect was the same for both feature (simple) and conjunction (complex) searches. Interestingly, in contrast to the compensatory effect of covert attention on information processing at iso-eccentric locations (temporal performance fields), covert attention did not eliminate speed differences across eccentricity.

Spatial resolution for feature binding is impaired in peripheral and amblyopic vision

We measured spatial resolution for discriminating targets that differed from nearby distractors in either color or orientation or their conjunction. In the fovea of normal human observers, whenever both attributes are big enough to be individually visible, their conjunction is also visible. In the periphery, the two attributes may be visible, but their conjunction may be invisible. We found a similar impairment in resolving conjunctions for the fovea of deprived eyes of humans with abnormal visual development (amblyopia). These results are quantitatively explained by a model of primary visual cortex (V1) in which orientation and color maps are imperfectly co-registered topographically. Our results in persons with amblyopia indicate that the ability of the fovea to compensate for this poor co-registration is consolidated by visual experience during postnatal development.

Attention enhances contrast sensitivity at cued and impairs it at uncued locations

Transient covert attention increases contrast sensitivity at the target location with an informative spatial cue. Here we explored whether an uninformative spatial cue (50% valid with two possible locations) also increases contrast sensitivity and whether contrast sensitivity is altered at the uncued location as compared to the neutral condition. For all four observers, transient covert attention had both a benefit and a cost: it enhanced contrast sensitivity at the cued location and impaired contrast sensitivity at the uncued location at both parafoveal and peripheral positions. These results are consistent with the idea of limited resources, and indicate that transient attention helps control the expenditure of cortical computation.

Greater depth seen with phantom stereopsis is coded at the early stages of visual processing

A visual search task was used to investigate the spatially parallel coding of depth from binocular disparity and from binocularly unmatched features. Experiment 1, using disparity noise, showed that detectability is higher for illusory phantom targets defined by unmatched features than for disparity-defined targets, although the two targets were equated as to theoretically minimum depth. Experiment 2, using binocularly unmatched noise whose width was equal to the disparity of the noise used in Experiment 1, showed that noise severely interferes with the detection of both the disparity and the phantom targets. These results are consistent with the idea that the greater depth seen with phantom stereopsis is coded at the early stages of visual processing.

Optimal eye movement strategies in visual search

To perform visual search, humans, like many mammals, encode a large field of view with retinas having variable spatial resolution, and then use high-speed eye movements to direct the highest-resolution region, the fovea, towards potential target locations. Good search performance is essential for survival, and hence mammals may have evolved efficient strategies for selecting fixation locations. Here we address two questions: what are the optimal eye movement strategies for a foveated visual system faced with the problem of finding a target in a cluttered environment, and do humans employ optimal eye movement strategies during a search? We derive the ideal bayesian observer for search tasks in which a target is embedded at an unknown location within a random background that has the spectral characteristics of natural scenes. Our ideal searcher uses precise knowledge about the statistics of the scenes in which the target is embedded, and about its own visual system, to make eye movements that gain the most information about target location. We find that humans achieve nearly optimal search performance, even though humans integrate information poorly across fixations. Analysis of the ideal searcher reveals that there is little benefit from perfect integration across fixations--much more important is efficient processing of information on each fixation. Apparently, evolution has exploited this fact to achieve efficient eye movement strategies with minimal neural resources devoted to memory.

Mask-dependent attentional cuing effects in visual signal detection: the psychometric function for contrast

A spatial-cuing paradigm was used to test the hypothesis of Carrasco, Penpeci-Talgar, and Eckstein (2000) that the mask-dependent cuing effects found in visual signal detection by Smith (2000a) were caused by submaximal activation of the transient-orienting system. Mask-dependent cuing was found with a range of stimulus contrasts with pure peripheral cues and with the mixed central-peripheral cues of Smith (2000a), contrary to the predictions of the submaximal activation hypothesis. The use of a pedestal detection task to control spatial uncertainty showed that the cuing effect was due to signal enhancement. A model of mask-dependent cuing is described, which assumes that attention affects the rate of information accumulation from the display and that masks limit the visual persistence of the stimulus. The model correctly predicts differential mask dependencies in sensitivity for detection and discrimination and the associated patterns of response times.

Visual Search and the Collapse of Categorization

Categorization researchers typically present single objects to be categorized. But real-world categorization often involves object recognition within complex scenes. It is unknown how the processes of categorization stand up to visual complexity or why they fail facing it. The authors filled this research gap by blending the categorization and visual-search paradigms into a visual-search and categorization task in which participants searched for members of target categories in complex displays. Participants have enormous difficulty in this task. Despite intensive and ongoing category training, they detect targets at near-chance levels unless displays are extremely simple or target categories extremely focused. These results, discussed from the perspectives of categorization and visual search, might illuminate societally important instances of visual search (e.g., diagnostic medical screening).

Attention capacity and task difficulty in visual search

When a visual search task is very difficult (as when a small feature difference defines the target), even detection of a unique element may be substantially slowed by increases in display set size. This has been attributed to the influence of attentional capacity limits. We examined the influence of attentional capacity limits on three kinds of search task: difficult feature search (with a subtle featural difference), difficult conjunction search, and spatial-configuration search. In all 3 tasks, each trial contained sixteen items, divided into two eight-item sets. The two sets were presented either successively or simultaneously. Comparison of accuracy in successive versus simultaneous presentations revealed that attentional capacity limitations are present only in the case of spatial-configuration search. While the other two types of task were inefficient (as reflected in steep search slopes), no capacity limitations were evident. We conclude that the difficulty of a visual search task affects search efficiency but does not necessarily introduce attentional capacity limits.

Specific-Token Effects in Screening Tasks: Possible Implications for Aviation Security

Screeners at airport security checkpoints perform an important categorization task in which they search for threat items in complex x-ray images. But little is known about how the processes of categorization stand up to visual complexity. The authors filled this research gap with screening tasks in which participants searched for members of target categories in visual displays. The authors found that when targets were sampled with replacement and repetition, participant screeners relied on recognizing familiar targets and had great difficulty using category-general knowledge. The authors observed a "heartbeat" in detection performance--it improved while test images repeated but dropped sharply when unfamiliar targets from the same categories appeared. This reliance on familiarity illuminates the processes of categorization under conditions of visual complexity and suggests limits on those processes. This reliance also has implications for the training and evaluation of screeners in the field.

Attentional effects on sensory tuning for single-feature detection and double-feature conjunction

When humans scan their visual environment, relevant objects are selectively attended for enhanced processing. It is still unclear in what ways processing is modified by attention, and whether attentional selection operates on an individual feature (such as colour, orientation or motion) or on binding together different features. In the experiments reported in this paper, these two stages were characterized using psychophysical reverse correlation. Subjects viewed eight patches, briefly flashed and symmetrically arranged around fixation. Each patch consisted of segments that could vary in both colour and orientation. One of the patches ('target') differed from the remaining 'distractor' patches with respect to either its orientation, colour, or both (in three different experiments). Subjects were asked to detect the target patch. The stimulus was preceded by a cue. On some trials ('cued' trials), the cue informed observers that the target patch could only appear at two of the eight possible locations. On remaining ('uncued') trials, all eight positions were valid. Psychophysical reverse correlation was then applied to derive linear estimates of sensory filters for orientation only, colour only, and their conjunction. In line with the properties of single neurons in cortex, attentional cueing did not affect sensory tuning for detecting individual features. However, it affected the way in which features were subsequently (and very inefficiently) combined in a multiplicative fashion. The results are consistent with a model in which attention recalibrates internal responses to the statistics of the stimulus by having signals from different features mutually control each other through reciprocal inhibition.

Are there capacity limitations in symmetry perception?

Previous researchers have proposed that there are two types of symmetry detection: one based on crude preattentive symmetry judgments and another based on detailed scrutiny of individual parts (Barlow & Reeves, 1979; S. E. Palmer & Hemenway, 1978; Royer, 1981). Four experiments were conducted to examine capacity limits in different symmetry judgments. Observers were required to discriminate between random patterns and approximate symmetry (Experiments 1 and 3) or between perfect and approximate symmetry (Experiments 2 and 4). The patterns were divided into two sets of dots, presented either simultaneously or successively. A comparison of accuracy under these two presentation conditions suggested that symmetry detection involves an analysis that is spatially parallel but coarse, regardless of either task difficulty or task type (detecting symmetry vs. detecting asymmetry).

The footprints of visual attention during search with 100% valid and 100% invalid cues

Human performance during visual search typically improves when spatial cues indicate the possible target locations. In many instances, the performance improvement is quantitatively predicted by a Bayesian or quasi-Bayesian observer in which visual attention simply selects the information at the cued locations without changing the quality of processing or sensitivity and ignores the information at the uncued locations. Aside from the general good agreement between the effect of the cue on model and human performance, there has been little independent confirmation that humans are effectively selecting the relevant information. In this study, we used the classification image technique to assess the effectiveness of spatial cues in the attentional selection of relevant locations and suppression of irrelevant locations indicated by spatial cues. Observers searched for a bright target among dimmer distractors that might appear (with 50% probability) in one of eight locations in visual white noise. The possible target location was indicated using a 100% valid box cue or seven 100% invalid box cues in which the only potential target locations was uncued. For both conditions, we found statistically significant perceptual templates shaped as differences of Gaussians at the relevant locations with no perceptual templates at the irrelevant locations. We did not find statistical significant differences between the shapes of the inferred perceptual templates for the 100% valid and 100% invalid cues conditions. The results confirm the idea that during search visual attention allows the observer to effectively select relevant information and ignore irrelevant information. The results for the 100% invalid cues condition suggests that the selection process is not drawn automatically to the cue but can be under the observers' voluntary control.

The roles of distractor noise and target certainty in search: a signal detection model

Observers searched arrays of briefly presented near-isoluminant colored disks for a single disk of known color (feature search) or unknown color (oddity search). Speed and accuracy were converted to a single, model-based measure of performance (Perf), in units of (d')2 per second of latency. Perf decreased with set size in feature search and increased in oddity. In both types of search, grouping the distractors, making them homogeneous in color, and reducing their saturation, all increased Perf. These commonalities suggested an SDT-based model in which distractors increase noise in the same way in both types of search. However, in oddity, though not in feature search, distractors must be attended and so adding distractors also boosts the effective target contrast, overcoming the added noise. A model with two free parameters for noise and one for attention accounted for every combination except for oddity searches among heterogeneous grouped distractors.

Temporal performance fields: visual and attentional factors

This study is the first to investigate: (a) 'temporal performance fields,' whether the speed of information accrual differs for different locations at a fixed eccentricity, and (b) whether covert attention modulates temporal dynamics differentially at isoeccentric locations. Using the speed accuracy tradeoff (SAT) procedure, we derived conjoint measures of how isoeccentric locations and precueing targets location affect speed and accuracy in a search task. The results demonstrate the existence of temporal performance fields, analogous to spatial performance fields: information accrual was fastest for target on the horizontal meridian, intermediate for targets at the intercardinal locations, slow for targets on the vertical meridian, and slowest for targets at the North (N) location (accrual time pattern: E&W<intercardinal<S<N). Surprisingly, in contrast to spatial performance fields, where covert attention enhanced discriminability at all locations to a similar degree, attention differentially sped up processing at the slower locations, with a greater benefit evident along the vertical than the horizontal meridian, particularly at the N location, and an intermediate benefit at intercardinal locations (viz., N>S>intercardinal>E&W). Hence, the compensatory effect of attention eliminated the temporal asymmetries across isoeccentric locations.

Temporal tuning characteristics of the perceptual template and endogenous cuing of spatial attention

External noise presented in temporal contiguity with a target impairs perceptual performance, reflecting the temporal tuning of the perceptual template. Deriving the temporal characteristics of the perceptual template, however, requires an observer model that segregates the impact of non-linearities and intrinsic inefficiencies of the observer in order to account for the impact of external noise in various temporal configurations. In Experiment 1, we showed that the perceptual template model successfully accounts for temporal masking functions in foveal Gabor orientation identification masked by external noise with a wide range of temporal configurations, and estimates the temporal characteristics of the perceptual template. In Experiment 2, we extended the paradigm and the model to compare the temporal tuning characteristics of the perceptual template in central pre- and simultaneous cuing of endogenous spatial attention in a Gabor orientation identification task in visual periphery. We found that endogenous spatial attention excludes external noise by both sharpening the temporal window of the perceptual template and (mostly) reducing the impact of external noise uniformly across the entire temporal window.

Attention orienting and the time course of perceptual decisions: response time distributions with masked and unmasked displays

Mask-dependent cuing effects, like those previously found in yes-no detection, were found in a task in which observers judged the orientations of orthogonally-oriented Gabor patches presented at cued or uncued locations. Attentional cues enhanced sensitivity for masked, but not unmasked, stimuli. Responses were faster to cued than to uncued stimuli, irrespective of masking. The distributions of response times and accuracy were well described by a diffusion process model of decision making. Mask-dependent cuing was explained by an orienting model in which: (a) decisions are based on stable stimulus representations in visual short term memory that determine the rate of evidence accumulation in the diffusion process; (b) inattention delays the entry of stimuli into short term memory, and (c) masks limit the visual persistence of stimuli.

A dynamic model of how feature cues guide spatial attention

We will describe a computational model of attention which explains the guidance of spatial attention by feedback within a distributed network. We hypothesize that feedback within the ventral pathway transfers the target template from prefrontal areas into intermediate areas like V4. The oculomotor circuit consisting of FEF, LIP and superior colliculus picks up this distributed activity and provides a continuous spatial reentry signal from premotor cells. In order to test this hypothesis, we simulate two experiments that require a response given a color cue. The first experiment indicates a parallel feature-based enhancement prior to any spatial selection. If two targets are behaviorally relevant, as in the second experiment, experimental findings indicate that subjects split their attention between two locations containing the searched feature. Our simulation results suggest that the split in attention between two foci is a transient effect occurring during competition. We predict that the time after cue presentation determines the state of this competition and ultimately the distribution of attention at different locations. In addition we provide simulation results to explain how reentrant processing through the oculomotor circuit might lead to variations of the time for target detection in visual search.

Object perception as Bayesian inference

We perceive the shapes and material properties of objects quickly and reliably despite the complexity and objective ambiguities of natural images. Typical images are highly complex because they consist of many objects embedded in background clutter. Moreover, the image features of an object are extremely variable and ambiguous owing to the effects of projection, occlusion, background clutter, and illumination. The very success of everyday vision implies neural mechanisms, yet to be understood, that discount irrelevant information and organize ambiguous or noisy local image features into objects and surfaces. Recent work in Bayesian theories of visual perception has shown how complexity may be managed and ambiguity resolved through the task-dependent, probabilistic integration of prior object knowledge with image features.

Cumulative Progress in Formal Theories of Attention

Formal theories of attention based on similarity-choice theory and signal-detection theory are reviewed to document cumulative progress in theoretical understanding of attention from the 1950s to the present. Theories based on these models have been developed to account for a wide variety of attentional phenomena, including attention to dimensions, attention to objects, and executive control. The review describes the classical similarity-choice and signal-detection theories and relates them to current theories of categorization, Garner tasks, visual search, cuing procedures, task switching, and strategy choice.

The multi-item localization (MILO) task: Measuring the spatiotemporal context of vision for action

We describe a new multi-item localization task that can be used to probe the temporal and spatial contexts of search-like behaviors. A sequence of four target letters (e.g., E, F, G, and H) was presented among four distractor letters. Observers located the targets in order. Both retrospective and prospective components of performance were examined. The retrospective component was assessed by having target items either vanish or remain once they had been located. This manipulation had little effect on search performance, suggesting that old target items can be efficiently ignored. The prospective component was assessed by shuffling future target and distractor locations after each response. This manipulation revealed that observers typically plan ahead at least one target into the future. However, even when observers cannot plan ahead, they are still able to ignore old targets. These findings suggest that both "what you did" and "what you intend to do" can influence the localization and selection of targets.

An ideal observer with channels versus feature-independent processing of spatial frequency and orientation in visual search performance

An influential assumption for the front end of models in vision, visual search, and object recognition is an analysis of independent features that correspond to basic image properties, such as motion, shape, and color. Empirically, one common test of independent features (a cue-summation study) measures performance with increasing available cues or features, with improving performance leading to conclusions of summation across independent features. In a study by Shimozaki et al. [J. Vision 2, 354-370 (2002)], both ideal and human observers showed no summation with large stimulus differences, in contrast to independent-feature models and suggesting that stimulus information (as assessed by an ideal observer) might affect cue-summation studies. Extending the previous summation study, observers performed a visual search of four Gabors differing in only orientation, only spatial frequency, or both orientation and spatial frequency, across a range of target-distractor differences. An ideal observer underpredicted human summation for small differences, whereas the independent-orientation and spatial-frequency feature models overpredicted human summation for large differences. An ideal observer with channels jointly tuned to spatial frequency and orientation predicted human performance across both small and large target-distractor differences.

Visual feature integration theory: past, present, and future

Visual feature integration theory was one of the most influential theories of visual information processing in the last quarter of the 20th century. This article provides an exposition of the theory and a review of the associated data. In the past much emphasis has been placed on how the theory explains performance in various visual search tasks. The relevant literature is discussed and alternative accounts are described. Amendments to the theory are also set out. Many other issues concerning internal processes and representations implicated by the theory are reviewed. The article closes with a synopsis of what has been learned from consideration of the theory, and it is concluded that some of the issues may remain intractable unless appropriate neuroscientific investigations are carried out.

Spatial attention: normal processes and their breakdown

Although "attention" is a general term in everyday folk and psychologic use, using research from cognitive psychology allows a focus on the processes associated with attention. A process-oriented definition of attention [2] makes "attention" a concept that can be studied rigorously. Attention is necessary for eliminating unwanted sensory inputs or irrelevant behavioral tasks and is useful when some cognitive system or process receives too many inputs. Attention acts to restrict the number of inputs and allow processing to continue in an effective manner. Although there are many forms of attentional selection, spatial attention is the most studied and perhaps the best understood form of selection. Spatial attention is the variety of attention most widely studied in neuropsychologic populations. As was evident from the authors' review, different neuropsychologic syndromes can be characterized as involving different difficulties with the component processes of spatial attention. Different neural structures work in concert to produce normal spatial selection, and damage to any of these neural structures, including the parietal and frontal lobes, pulvinar, and superior colliculus, produce an attentional impairment. Attentional impairments often have been studied in cases of focal brain damage resulting from stroke, traumatic brain injury, tumor, and surgical resection. Focal or multifocal lesions caused by neurodegenerative impairments, such as PSP and AD, however, also have marked effects on neural processes of attention. Patients with AD exhibit a range of impairments in spatial attention, and these impairments have been studied relatively little. Despite the understanding of attention provided by various patient populations, gaps remain in knowledge of the mechanisms involved in attending to space. For example, most studies of neurologic patients have used a simple spatial cuing paradigm with highly predictive cues. Although this task yields robust results and can be adapted for use with many different patient groups, the spatial cuing effects produced by such tasks have multiple interpretations (e.g., allocation of resources versus reduction of decision noise). This review highlights the general point that different neural sites seem to be responsible for different forms of attentional control, such as object-based attention, attentional task switching, and the executive control of attention [8.33]. Understanding the integration of these neural sites and their relationship to cognitive processes and, ultimately, behavior, will increase understanding of normal and disordered attentional selection.

Effect of parietal lobe lesions on saccade targeting and spatial memory in a naturalistic visual search task

The eye movements of two patients with parietal lobe lesions and four normal observers were measured while they performed a visual search task with naturalistic objects. Patients were slower to perform the task than the normal observers, and the patients had more fixations per trial, longer latencies for the first saccade during the visual search, and less accurate first and second saccades to the target locations during the visual search. The increases in response times for the patients compared to the normal observers were best predicted by increases in the number of fixations. In order to investigate the effects of spatial memory on search performance, in some trials observers saw a preview of the search display. The patients appeared to have difficulty using previously viewed information, unlike normal observers who benefit from the preview. This suggests a spatial memory deficit. The patients' deficits are consistent with the hypothesis that the parietal cortex has a role in the selection of targets for saccades, in memory for target location.

Bayesian models of object perception

The human visual system is the most complex pattern recognition device known. In ways that are yet to be fully understood, the visual cortex arrives at a simple and unambiguous interpretation of data from the retinal image that is useful for the decisions and actions of everyday life. Recent advances in Bayesian models of computer vision and in the measurement and modeling of natural image statistics are providing the tools to test and constrain theories of human object perception. In turn, these theories are having an impact on the interpretation of cortical function.

Does partial difficult search help difficult search?

Olds, Cowan, and Jolicoeur (2000a, 2000b) showed that exposure to a display that affords pop-out search a target among distractors of only one color) can assist processing of a related display that requires difficult search. They added distractors of an additional color to the initial simple display and analyzed response time distributions to show that exposure to the initial display aided subsequent search in the difficult portion (this finding was called search assistance). To test whether search assistance depends on perceptual grouping of the initial items, we presented initial items that were more difficult to group (two colors of distractors, instead of just one). The target appeared (on 50% of the trials) among distractors of two colors, and then after a delay, more distractors of those two colors were added to the display. Exposure to the initial easier portion of the display did not assist processing of the second portion of the display when the initial display contained a large number of items; we found tentative evidence for assistance with small numbers of initial items. In the Olds et al. (2000a, 2000b) studies, it was easy to group the initial distractor items, because they were all the same color. In contrast, in the present study, it was difficult to group the heterogeneous initial distractor items. Search assistance is found only when initial item grouping is relatively easy, and thus we conclude that search assistance depends on grouping.