Reconciling change blindness with long-term memory for objects

How can we reconcile remarkably precise long-term memory for thousands of images with failures to detect changes to similar images? We explored whether people can use detailed, long-term memory to improve change detection performance. Subjects studied a set of images of objects and then performed recognition and change detection tasks with those images. Recognition memory performance exceeded change detection performance, even when a single familiar object in the postchange display consistently indicated the change location. In fact, participants were no better when a familiar object predicted the change location than when the displays consisted of unfamiliar objects. When given an explicit strategy to search for a familiar object as a way to improve performance on the change detection task, they performed no better than in a 6-alternative recognition memory task. Subjects only benefited from the presence of familiar objects in the change detection task when they had more time to view the prechange array before it switched. Once the cost to using the change detection information decreased, subjects made use of it in conjunction with memory to boost performance on the familiar-item change detection task. This suggests that even useful information will go unused if it is sufficiently difficult to extract.     

Change detection in the flicker paradigm: Do faces have an advantage?

Changes to a scene that occur during an eye movement, image flicker, or movie cut are difficult to detect. One way to measure change detection performance is with the flicker paradigm, where changes between two images are introduced during a brief blank screen, which causes the images to “flicker”. Ro, Russell, and Lavie (2001) presented flickering displays consisting of one face and five different common objects. They found that changes to faces were detected both more rapidly and more accurately than changes to objects and suggested that faces capture attention due to their biological significance. In the present studies, we found that changes to objects were more readily detected than changes to faces when displays consisted of an object among a number of faces. That is, a change detection advantage was observed for the “odd-one-out” in the array, regardless of its significance. Therefore, faces may not have a special status for change detection.     

Object and observer motion in the perception of objects by infants

Sixteen-week-old human infants distinguish optical displacements given by their own motion from displacements given by moving objects, and they use only the latter to perceive the unity of partly occluded objects. Optical changes produced by moving the observer around a stationary object produced attentional levels characteristic of stationary observers viewing stationary displays and much lower than those shown by stationary observers viewing moving displays. Real displacements of an object with no subject-relative displacement, produced by moving an object so as to maintain a constant relation to the moving observer, evoked attentional levels that were higher than with stationary displays and more characteristic of attention to moving displays, a finding suggesting detection of the real motion. Previously reported abilities of infants to perceive the unity of partly occluded objects from motion information were found to depend on real object motion rather than on optical displacements in general. The results suggest that object perception depends on registration of the motions of surfaces in the three-dimensional layout.     

The relationship between change detection and recognition of centrally attended objects in motion pictures

Observers typically detect changes to central objects more readily than changes to marginal objects, but they sometimes miss changes to central, attended objects as well. However, even if observers do not report such changes, they may be able to recognize the changed object. In three experiments we explored change detection and recognition memory for several types of changes to central objects in motion pictures. Observers who failed to detect a change still performed at above chance levels on a recognition task in almost all conditions. In addition, observers who detected the change were no more accurate in their recognition than those who did not detect the change. Despite large differences in the detectability of changes across conditions, those observers who missed the change did not vary in their ability to recognize the changing object.     

Change blindness and visual memory: Visual representations get rich and act poor

Change blindness is often taken as evidence that visual representations are impoverished, while successful recognition of specific objects is taken as evidence that they are richly detailed. In the current experiments, participants performed cover tasks that required each object in a display to be attended. Change detection trials were unexpectedly introduced and surprise recognition tests were given for nonchanging displays. For both change detection and recognition, participants had to distinguish objects from the same basic‐level category, making it likely that specific visual information had to be used for successful performance. Although recognition was above chance, incidental change detection usually remained at floor. These results help reconcile demonstrations of poor change detection with demonstrations of good memory because they suggest that the capability to store visual information in memory is not reflected by the visual system's tendency to utilize these representations for purposes of detecting unexpected changes.     

Why don't we see changes?: The role of attentional bottlenecks and limited visual memory

Seven experiments explore the role of bottlenecks in selective attention and access to visual short-term memory in the failure of observers to identify clearly visible changes in otherwise stable visual displays. Experiment One shows that observers fail to register a color change in an object even if they are cued to the location of the object by a transient at that location as the change is occurring. Experiment Two shows the same for orientation change. In Experiments Three and Four, attention is directed to specific objects prior to making changes in those objects. Observers have only a very limited memory for the status of recently attended items. Experiment Five reveals that observers have no ability to detect changes that happen after attention has been directed to an object and before attention returns to that object. In Experiment Six, attention is cued at rates that more closely resemble natural rates and Experiment Seven uses natural images. Memory capacity remains very small (<4 items).     

Active and passive scene recognition across views.

Recent evidence suggests that scene recognition across views is impaired when an array of objects rotates relative to a stationary observer, but not when the observer moves relative to a stationary display [Simons, D.J., Wang, R.F., 1998. Perceiving real-world viewpoint changes. Psychological Science 9, 315-320]. The experiments in this report examine whether the relatively poorer performance by stationary observers across view changes results from a lack of perceptual information for the rotation or from the lack of active control of the perspective change, both of which are present for viewpoint changes. Three experiments compared performance when observers passively experienced the view change and when they actively caused the change. Even with visual information and active control over the display rotation, change detection performance was still worse for orientation changes than for viewpoint changes. These findings suggest that observers can update a viewer-centered representation of a scene when they move to a different viewing position, but such updating does not occur during display rotations even with visual and motor information for the magnitude of the change. This experimental approach, using arrays of real objects rather than computer displays of isolated individual objects, can shed light on mechanisms that allow accurate recognition despite changes in the observer's position and orientation.     

Effects of spatial configurations on visual change detection: An account of bias changes

In order to determine whether people encode spatial configuration information when encoding visual displays, in four experiments, we investigated whether changes in task-irrelevant spatial configuration information would influence color change detection accuracy. In a change detection task, when objects in the test display were presented in new random locations, rather than identical or different locations preserving the overall configuration, participants were more likely to report that the colors had changed. This consistent bias across four experiments suggested that people encode task-irrelevant spatial configuration along with object information. Experiment 4 also demonstrated that only a low-false-alarm group of participants effectively bound spatial configuration information to object information, suggesting that these types of binding processes are open to strategic influences.     

A temporal same-object advantage in the tunnel effect: facilitated change detection for persisting objects

Meaningful visual experience requires computations that identify objects as the same persisting individuals over time, motion, occlusion, and featural change. This article explores these computations in the tunnel effect: When an object moves behind an occluder, and then an object later emerges following a consistent trajectory, observers irresistibly perceive a persisting object, even when the pre- and postocclusion views contrast featurally. This article introduces a new change detection method for quantifying percepts of the tunnel effect. Observers had to detect color changes in displays where several objects oscillated behind occluders and occasionally changed color. Across comparisons with several types of spatiotemporal gaps, as well as manipulations of occlusion versus implosion, performance was better when objects' kinematics gave the impression of a persisting individual. The results reveal a temporal same-object advantage: better change detection across temporal scene fragments bound into the same persisting object representations. This suggests that persisting objects are the underlying units of visual memory.     

Perceptual expertise improves category detection in natural scenes

There is much debate about how detection, categorization, and within-category identification relate to one another during object recognition. Whether these tasks rely on partially shared perceptual mechanisms may be determined by testing whether training on one of these tasks facilitates performance on another. In the present study we asked whether expertise in discriminating objects improves the detection of these objects in naturalistic scenes. Self-proclaimed car experts (N = 34) performed a car discrimination task to establish their level of expertise, followed by a visual search task where they were asked to detect cars and people in hundreds of photographs of natural scenes. Results revealed that expertise in discriminating cars was strongly correlated with car detection accuracy. This effect was specific to objects of expertise, as there was no influence of car expertise on person detection. These results indicate a close link between object discrimination and object detection performance, which we interpret as reflecting partially shared perceptual mechanisms and neural representations underlying these tasks: the increased sensitivity of the visual system for objects of expertise – as a result of extensive discrimination training – may benefit both the discrimination and the detection of these objects. Alternative interpretations are also discussed.     

The role of attention in processing configural and shape information in 3-D novel objects

Recent research suggests that there is an advantage for processing configural information in scenes and objects. The purpose of this study was to investigate the extent to which attention may account for this configural advantage. In Experiment 1, we found that cueing the location of change in single object displays improved detection performance for both configural and shape changes, yet cueing attention away from the location of change was detrimental only for shape change detection. A configural advantage was present for each cueing condition. Experiments 2A and 2B examined whether the configural advantage persisted in conditions where attention was distributed more widely, using a visual search paradigm. Although searches for configural changes were faster than those for shape changes across all set sizes, both types of information appeared to be processed with similar efficiency. Overall, these results suggest that the configural advantage is independent of the location or distribution of visual attention.     

Visual short-term memory operates more efficiently on boundary features than on surface features

A change detection task was used to estimate the visual short-term memory storage capacity for either the orientation or the size of objects. On each trial, several,objects were briefly presented, followed by a blank interval and then by a second display of objects that either was identical to the first display or had a single object that was different (the object changed either orientation or size, in separate experiments). The task was to indicate whether the two displays were the same or different, and the number of objects remembered was estimated from the percent correct on this task. Storage capacity for a feature was nearly twice as large when that feature was defined by the object boundary, rather than by the surface texture of the object. This dramatic difference in storage capacity suggests that a particular feature (e.g., right tilted or small) is not stored in memory with an invariant abstract code. Instead, there appear to be different codes for the boundary and surface features of objects, and memory operates on boundary features more efficiently than it operates on surface features.     

Object attention revisited: identifying mechanisms and boundary conditions

Multiple attributes of a visual array are often more efficiently processed when they are attributes of a single object than when they are attributes of different objects—a pattern reflecting the limitations of object attention. This study used psychophysical methods to evaluate the object attention limitations in the report of attributes (orientation and phase) computed early in visual analysis for spatially separated objects. These limitations had large effects on dual-object report thresholds when different judgments were required for the two objects (orientation for one object and phase for the other), but the effects were small or nonexistent when the same judgment was made about both objects. Judgment consistency reduced or eliminated the expression of object attention deficits. Thus, the deficits in dual-object report reflect both division of attention over objects and the calculation of independent reference or judgment operations. Dual-object deficits, when they occurred, were substantial in displays with external noise masks. Smaller effects were observed in clear displays, even when difficulty was equated by stimulus contrast. Thus, the primary consequence of object attention is the exclusion of external noise, or mask suppression, and enhancement of the stimulus in clear displays is a secondary consequence.     

Failure to detect changes to people during a real-world interaction

Recent research on change detection has documented surprising failures to detect visual changes occurring between views of a scene, suggesting the possibility that visual representations contain few details. Although these studies convincingly demonstrate change blindness for objects in still images and motion pictures, they may not adequately assess the capacity to represent objects in the real world. Here we examine and reject the possibility that change blindness in previous studies resulted from passive viewing of 2-D displays. In one experiment, an experimenter initiated a conversation with a pedestrian, and during the interaction, he was surreptitiously replaced by a different experimenter. Only half of the pedestrians detected the change. Furthermore, successful detection depended on social group membership; pedestrians from the same social group as the experimenters detected the change but those from a different social group did not. A second experiment further examined the importance of this effect of social group. Provided that the meaning of the scene is unchanged, changes to attended objects can escape detection even when they occur during a natural, real-world interaction. The discussion provides a set of guidelines and suggestions for future research on change blindness.     

Object-based warping: an illusory distortion of space within objects

Visual objects are high-level primitives that are fundamental to numerous perceptual functions, such as guidance of attention. We report that objects warp visual perception of space in such a way that spatial distances within objects appear to be larger than spatial distances in ground regions. When two dots were placed inside a rectangular object, they appeared farther apart from one another than two dots with identical spacing outside of the object. To investigate whether this effect was object based, we measured the distortion while manipulating the structure surrounding the dots. Object displays were constructed with a single object, multiple objects, a partially occluded object, and an illusory object. Nonobject displays were constructed to be comparable to object displays in low-level visual attributes. In all cases, the object displays resulted in a more powerful distortion of spatial perception than comparable non-object-based displays. These results suggest that perception of space within objects is warped.     

Detecting changes between real-world objects using spatiochromatic filters

A behavioral and computational treatment of change detection is reported. The behavioral task was to judge whether a single object substitution change occurred between two “flickering” 9-object scenes. Detection performance was found to vary with the similarity of the changing objects; object changes violating orientation and category yielded the fastest and most accurate detection responses. To account for these data, theBOLAR model was developed, which uses color, orientation, and scale selective filters to compute the visual dissimilarity between the pre- and postchange objects from the behavioral study. Relating the magnitude of the BOLAR difference signals to change detection performance revealed that object pairs estimated as visually least similar were the same object pairs most easily detected by observers. The BOLAR model advances change detection theory by (1) demonstrating that the visual similarity between the change patterns can account for much of the variability in change detection behavior, and (2) providing a computational technique for quantifying these visual similarity relationships for real-world objects.     

Eye movements during change detection: implications for search constraints, memory limitations, and scanning strategies

Search, memory, and strategy constraints on change detection were analyzed in terms of oculomotor variables. Observers viewed a repeating sequence of three displays (Scene 1-->Mask-->Scene 2-->Mask...) and indicated the presence-absence of a changing object between Scenes 1 and 2. Scenes depicted real-world objects arranged on a surface. Manipulations included set size (one, three, or nine items) and the orientation of the changing objects (similar or different). Eye movements increased with the number of potentially changing objects in the scene, with this set size effect suggesting a relationship between change detection and search. A preferential fixation analysis determined that memory constraints are better described by the operation comparing the pre- and postchange objects than as a capacity limitation, and a scanpath analysis revealed a change detection strategy relying on the peripheral encoding and comparison of display items. These findings support a signal-in-noise interpretation of change detection in which the signal varies with the similarity of the changing objects and the noise is determined by the distractor objects and scene background.     

Phonological similarity effects on detecting change in simple arrays

In two experiments, we investigated the effects on change detection and identification of the phonological similarity of the names of drawn objects shown in simple arrays. Experiment 1 examined the effects of phonological similarity on report of identity changes (in which one object was replaced by another) in a five-item array and Experiment 2 its effects on reporting switch changes (in which objects exchanged places). Phonological similarity did not impair change detection, and identification of new objects on change trials was more accurate in Experiment 1. H owever, phonological similarity significantly reduced change detection and identification of objects that switched location in Experiment 2. The results indicate different roles for a visual short-term memory system such as the visuospatial sketchpad and for phonological coding and storage in detecting and identifying replaced objects and object locations in arrays.     

Category-based grouping in working memory and multiple object tracking

Two prominent cognitive capacity limitations are the maximal number of objects we can place in working memory (WM) and the maximal number of objects we can track in a display. Both are believed to have a numeric value of 3 or 4, which has led to the proposal that we have a general cognitive capacity, and that this capacity is most likely linked to limitations of how many objects we can attend simultaneously. Based on previous results showing that we can memorize more objects if they come from different categories than if they come from the same category (e.g., Feigenson & Halberda, 2008; Wong, Peterson, & Thompson, 2008; Wood, 2008), we compare how category-based grouping affects performance for WM and multiple object tracking (MOT). We present participants with either “pure” displays of either cars or faces, or with “mixed” displays of cars and faces. Overall, the effects of category are weak. In some analyses but not others, we replicate the mixed advantage for WM, albeit with a small effect size. In contrast, we observe a weak pure advantage for MOT tasks, at least in a meta-analysis of five experiments, but not in all experiments. Accordingly, WM and MOT tasks differed significantly in their sensitivity to category membership. We also find that WM is slightly better for faces than for cars, but that no such difference exists for MOT. We tentatively suggest that cognitive capacity limitations in different domains are at least partially due to the limitations of distinct mechanisms.     

Detection and Identification of Change in Naturalistic Scenes

Research using change detection paradigms has demonstrated that only limited scene information remains available for conscious report following initial inspection of a scene. Previous researchers have found higher change identification rates for deletions of parts of objects in line drawings of scenes than additions. Other researchers, however, have found an asymmetry in the opposite direction for addition/deletion of whole objects in line drawings of scenes. Experiment 1 investigated subjects' accuracy in detecting and identifying changes made to successive views of high quality photographs of naturalistic scenes that involved the addition and deletion of objects, colour changes to objects, and changes to the spatial location of objects. Identification accuracy for deletions from scenes was highest, with lower identification rates for object additions and colour changes, and the lowest rates for identification of location changes. Data further suggested that change identification rates for the presence/absence of objects were a function of the number of identical items present in the scene. Experiment 2 examined this possibility further, and also investigated whether the higher identification rates for deletions found in Experiment 1 were found for changes involving whole objects or parts of objects. Results showed higher identification rates for deletions, but only where a unique object was deleted from a scene. The presence of an identical object in the scene abolished this deletion identification advantage. Results further showed that the deletion/addition asymmetry occurs both when the objects are parts of a larger object and when they are entire objects in the scene.