Depending on their dominancy, availability, and validity, these multiple representations interact to determine memory performance. Specifically, representations that are automatically encoded and extensively practiced are more dominant, and their availability improves performance when they are valid. On the other hand, when the dominant representations are available but invalid, people may have to resort to the less dominant representations. As a result, the availability of these dominant but invalid representations can actually hurt performance, due to interference.
If these interfering representations are eliminated, the performance is again improved. The implications of these findings for general human spatial cognition are discussed. Download to read the full article text. Brewer, B. Frames of reference. Eilan, R. Brewer Eds. Oxford: Blackwell. Google Scholar.
Burgess, N. The hippocampal and parietal foundations of spatial cognition. New York: Oxford University Press. Integrating hippocampal and parietal functions: A spatial point of view.
Burgess, K. Colby, C. Space and attention in parietal cortex. Annual Review of Neuroscience , 22 , — PubMed Article Google Scholar. Diwadkar, V. Viewpoint dependence in scene recognition. Psychological Science , 8 , — Article Google Scholar. Driver, J. Object-centered visual neglect, or relative egocentric neglect? Journal of Cognitive Neuroscience , 12 , — Easton, R. Object-array structure, frames of reference, and retrieval of spatial knowledge. Egeth, H. Visual attention: Control, representation, and time course.
Annual Review of Psychology , 48 , — Franklin, N. Switching points of view in spatial mental models. Gallistel, C. The organization of learning. Garnham, A. A unified theory of the meaning of some spatial relational terms.
Cognition , 31 , 45— Halligan, P. Spatial cognition: Evidence from visual neglect. Trends in Cognitive Sciences , 7 , — Itti, L. Computational modeling of visual attention.
Nature Reviews , 2 , — Johnsrude, I. A cognitive activation study of memory for spatial relationships. Neuropsychologia , 37 , — Keppel, G. King, J. Human hippocampus and viewpoint dependence in spatial memory.
Hippocampus , 12 , — Klatzky, R. Allocentric and egocentric spatial representations: Definitions, distinctions, and interconnections. Freksa, C. Wender Eds. New York: Springer. Lansdale, M. Modeling memory for absolute location. Psychological Review , , — Levinson, S. Bloom, M. Eye tracking was not performed and chinrest was not used in this experiment. Because this experiment was different to the previous one with respect to several factors, e.
Instead, we concentrate only on qualitative differences and similarities. Note that in Experiment 2 identification performance was not significantly influenced by delay. These results replicate our previous findings that additional objects in memory lead to steeper degradation of localization performance. Similarly to experiment 2, when we controlled localization performance for swap errors by analyzing the distance between the reported location of the object and the nearest original location of any other object green line in Fig.
Thus, these results further strengthen the conclusion that increasing localization error with time due to having multiple objects to remember as manifested in the interaction between object-number and delay is largely due to swap errors.
Note that only trials in which objects were correctly identified were entered into the analysis. Therefore swapped objects are unlikely to be a result of failure to remember object identity. Nevertheless, some of the correct identifications are expected to occur by chance, and assuming location memory is intact, participants are expected to localize the items randomly in one of the remembered locations. In one out of three of such cases the object will be localized near its original location and in the remaining two thirds, around the other objects.
Similarly to the analysis in Experiment 2, we estimated the upper limit on the number of swap errors that could be attributed to guessing the correct identity by multiplying the percentages of the failed identification with two thirds grey line Figure 6D. We conclude that despite the various differences in the experimental settings e.
This provides important support for the conclusions that only 3 seconds of additional delay decrease localization performance, especially when several objects have to be remembered.
This degradation is associated with an increased probability to localize an item specifically at the location of another item in the memory array. In our experiments, participants were required to re-locate objects to their remembered locations. More objects in memory, as well as longer retention intervals, led to a progressive increase in mean localization error. Our analysis suggests that this additional error was caused by mislocalization to the locations of other objects in the memory array, presumably due to the fragility of links between object identity and location.
Indeed, only trials with correct identification were actually included in our analysis. Furthermore, the number of swap errors significantly exceeded the higher limit on swap errors that could be expected from correct identifications by chance. Thus, we conclude that swap errors are most likely to arise from binding failure between object identity and location information.
It has been shown before that locating objects to their correct locations is more difficult than remembering their identity or locations alone [50] , [54] , suggesting that links between object identity and location are particularly fragile. However, these studies did not investigate the distribution of errors or,more importantly, did not manipulate delay duration. Therefore those studies could not directly address how object identity and location information is maintained, or indeed forgotten over different time intervals.
We found that extending the maintenance period by only 3 seconds led to significantly higher number of binding failures supporting the claim that resources are indeed required to bind visual features to locations in memory [59] and challenging the claim that objects are maintained as an integrated unit in memory and forgotten as entirety [40].
The episodic buffer is assumed to be a limited capacity storage system capable of holding bound objects, but not performing the binding [61]. Importantly, in our experiment, participants could not predict the delay duration at the time the memory array was presented so presumably visual processing and feature integration were identical in both delay conditions.
They need to be actively linked over time for veridical recall of which object was where. We make no claim here as to the manner in which locations are memorized on their own. Locations could be represented either relative to the current fixation point [62] , to each other [63] or relative to the scene and context [44] , [45] , [64]. Previous studies have investigated the role of location in short-term memory of object identity.
In change detection tasks, at short delays, changing the location of objects between stimuli and test impairs detection of change [63] , suggesting a close link between objects and their locations in memory. Interestingly, the effect of scrambling item locations was found to diminish at longer delays [65].
This finding would be consistent with our conclusion that the links between identity and location degrade with time when multiple items have to be remembered. Treisman and colleagues have demonstrated that when participants are presented with several objects and later required to report the different features belonging to one of them, they often made a specific kind of error, often called a conjunction error. Rather than erroneously reporting a random value, they often swap features belonging to different objects.
Such errors were suggested to be a result of insufficient attentional resources that are needed, according to the Feature Integration Theory FIT , to bind together distinct features. However, there have been alternative interpretations. One study reported that the frequency of feature binding errors across various conditions is better explained by uncertainty about the location of visual features than FIT [66].
A more controlled manipulation of attention, in their view, reveals that the availability of attention resources does not in fact influence the frequency of conjunction errors [35]. Instead, they argued that conjunction errors are affected by post-perceptual rather than perceptual processes. Our findings demonstrate specifically that degradation of information during maintenance in WM, often closely linked to attention, can contribute to apparent binding failures at the report stage.
The additional swap errors we report for longer retention intervals are clearly a result of post perceptual processes that relate to increasing uncertainty about the location of the objects in a highly specific manner: being biased towards the locations of other objects in the memory array.
This study complements and extends recent findings from a color matching task [67] , in which, following a brief delay, participants had to reproduce the color of an oriented bar from a prior array of several bars of different colors and orientations.
Bays, Catalano and Husain found that a significant amount of the variability in the response could be explained by mis-reporting the features of the wrong item in memory. Moreover, such errors greatly increased when additional items had to be remembered. Most interestingly, another study demonstrated that errors resulting from visual crowding are not purely random, but they are similarly biased towards the features of the distractor items [68].
To conclude, in support of the existence of distinct memory representation for location and identity of objects, we have found that extending the retention interval by only 3 seconds led to an increased probability to swap the correct location and identity of objects held in memory, in a manner that could not be explained by forgetting of object identity or location alone.
Such binding failures significantly contribute to rapid short-term forgetting as measured by the decline in localization performance across time. Thus, when objects are forgotten they do not disappear completely from memory, as previously claimed, but rather the links to their locations are gradually broken. Conceived and designed the experiments: YP MH. Analyzed the data: YP. Wrote the paper: YP MH. Browse Subject Areas? Click through the PLOS taxonomy to find articles in your field. Introduction Remembering the locations of objects in space is a crucial biological function.
Experiment 1 Methods Participants. Download: PPT. Figure 1. Mean localization error increases with more objects to remember. Results and Discussion First we analyzed the absolute distance between the original and reported locations of the objects Fig. Figure 2. Localization error with respect to selection and fixation order. Figure 3. Figure 4. Experiment 2 Although in Experiment 1 we allowed 1 second per object to be encoded, can we be sure that swap errors relate to memory maintenance rather than failures related to imperfect processing at the perceptual stage?
Figure 5. Experiment 2: Recall of object identity and object location over two different delays. Figure 6. Experiment 3: Recall of identity and location of fractals over two different delays. Methods The procedure of Experiment 2 was exactly the same as that of Experiment 1, with the following exceptions.
Experiment 3 In the previous experiments we used images of real complex objects. Methods The procedure was the same as that of Experiment 2, with the following exceptions. Results and Discussion Because this experiment was different to the previous one with respect to several factors, e.
Discussion In our experiments, participants were required to re-locate objects to their remembered locations. References 1. Hollingworth A, Franconeri SL Object correspondence across brief occlusion is established on the basis of both spatiotemporal and surface feature cues.
Cognition — View Article Google Scholar 2. The Visual World in Memory. Current issues in memory. Psychology Press, Vol. Perception — View Article Google Scholar 4. Looking at versus looking for objects in scenes.
View Article Google Scholar 5. View Article Google Scholar 6. Visual Memory. Oxford Series in Visual Cogntion. Oxford University Press, Vol. View Article Google Scholar 7. DiCarlo JJ, Maunsell JH Anterior inferotemporal neurons of monkeys engaged in object recognition can be highly sensitive to object retinal position. J Neurophysiol — View Article Google Scholar 8.
J Comp Neurol — View Article Google Scholar 9. Neuron — View Article Google Scholar J Neurophysiol 96— J Neurophysiol Cortex 14— Science — Nat Neurosci 3: 85— Nat Neurosci 8: — J Cogn Psychol — Trends Cogn Sci — Pylyshyn ZW Visual indexes, preconceptual objects, and situated vision. Pylyshyn Z Some puzzling findings in multiple object tracking: I. Following the competition hypothesis, we expected an effect of emotion only in the second experiment. Similarly, Experiment 1C was performed by associating neutral and positive pictures the latter of the same arousal level as negative pictures used in the previous experiment to the rectangles, in order to explore competition mechanisms also with positive valenced stimuli.
Further experiments were performed to better clarify which emotional feature affects spatial working memory performance in this paradigm.
More specifically, in Experiment 2 we examined the effect of arousal by comparing the level of performance, under competition conditions i. Since in all previous experiments emotional and neutral pictures used within the same trial to elicit competition differed in both valence and arousal, we performed a further experiment in which we manipulated the valence of pictures, keeping the arousal level constant.
To this purpose, in Experiment 3A we presented negative and neutral pictures with similar low levels of arousal, and in Experiment 3B positive and neutral pictures again with similar low levels of arousal. Since all the experiments performed to examine competition entailed the presentation of emotional versus neutral stimuli within the same trial, we designed a last experiment in order to verify the competition hypothesis when emotional stimuli with different valence positive and negative , but similar arousal levels, compete with one another within the same trial Experiment 4.
A total of females; age: All subjects gave written informed consent in accordance with the Declaration of Helsinki. All participants were Italian speakers and reported having normal or corrected-to-normal vision. To study the effect of the incidental presentation of emotional stimuli on VSWM performance, a modified version of the object relocation task Kessels et al.
Participants sat in front of a PC screen and an instruction slide was shown in which they received information about the object-relocation task, but not about the presentation of emotional pictures. Figure 1. A schematic representation of the experimental procedure used in all experiments. Mean valence and arousal values for the pictures used in each experiment are reported in Table 1 , and IAPS codes in the Appendix.
When participants felt ready, they pressed a button to begin the encoding phase which was signaled by the presentation of a cross in the center of the screen for ms. After ms, eight pictures selected from IAPS appeared one at a time superimposed on each rectangle. Each picture was presented for ms ISI: ms. Thus, the encoding phase lasted 11 s and ms. The test phase took place immediately after the end of the encoding phase. All the black rectangles appeared at the bottom of the screen and participants had to relocate them as accurately as possible, using the mouse.
Memory for object location was evaluated considering the distance between the original position and the closest relocated object. Long-term memory for incidentally learned pictures was evaluated 24 h later by a free recall task: participants were asked to verbally recall, by speaking the name of the objects depicted in the pictures they had seen on the previous day.
After the free recall test, pictures were again presented one at a time, for 7 s, on the screen. Participants were instructed to view each picture and to subjectively evaluate valence and arousal of each picture by using the Self-Assessment Manikin SAM. In Experiment 1A object positions were tagged by presenting negative pictures to one group Negative group and neutral pictures to a second group Neutral group , in a between-subject manipulation. In brief, participants 32 females and 8 males; age Since in this experiment a between- subject design was used, to further investigate if performance would be affected by basal differences in motor and spatial abilities of participants we performed a supplemental experiment in which a control condition followed the experimental trial with IAPS picture.
In this control condition, the procedure was identical to that previously described but the images associated to rectangles were built by scrambling pixels of different colors. In Experiment 1B, within a single trial, half of the object positions were tagged by the presentation of neutral pictures and the other half by negative pictures, thus allowing a within-subject manipulation. In brief, after participants 20 females and 8 males; age: The starting position of picture presentation was varied and each of the eight rectangles on the display could act as the starting position, thus yielding eight different configurations.
Further, for half of the participants the first item presented was an emotional picture negative in Experiments 2, 3, and 5; positive in Experiment 4 , followed by a neutral one, while for the other half, presentation started with a neutral picture followed by an emotional one, resulting in 16 different configurations. The order was counterbalanced across participants. In brief, after participants 19 females and 6 males; age: The order of picture presentation was randomized like in Experiment 2.
Participants 30 females and 19 males; age: They were randomly assigned to two different groups: i High arousal in which object positions were tagged by four negative pictures with high arousal values, and four neutral pictures; ii Low arousal in which object positions were tagged by four negative pictures with low arousal values, and four neutral pictures. The present experiment was designed to investigate the effect of emotional valence on VSWM performance.
Participants 11 females and 9 males; age: Participants 17 females and 4 males; age: The last experiment was designed to further investigate the effect of emotional valence on VSWM performance.
Eighteen participants 13 females; age: The procedure was identical to the one used previously but the pictures differed in order to have four negative and four neutral pictures with similar levels of arousal. In all experiments the displacement error was calculated as the distance expressed in pixel between the center of the originally positioned object and the center of the closest relocated object.
The first experiments were carried out in order to verify the competition hypothesis. In Experiment 1A object-positions were tagged either by neutral or negative pictures. Thus, the effect of emotion on VSWM was investigated as a between-subject factor. Memory for incidentally learned pictures was evaluated in a free recall test carried out 24 h after the object-relocation task.
In order to investigate whether a confounding effect of the individual differences in motor control or in spatial working memory ability could influence relocation performance, we carried out a supplemental experiment in which 3 h after the main task performed with negative and neutral IAPS pictures IAPS pictures condition , participants performed a further object relocation task in which object-positions were tagged by the presentation of pictures built by scrambling pixels of different colors Control condition.
The analysis revealed no significant effects of Group [Negative vs. Figure 2. Effect of emotional stimuli on VSWM when neutral and emotional information is not A or is B,C presented within the same encoding trial.
Mean displacement error pixel displayed by participants in the immediate re-location test A when neutral and negative information is not presented within the same encoding trial, B when neutral and negative information is presented within the same encoding trial, and C when neutral and positive information is presented within the same encoding trial.
In Experiment 1B, object positions during the encoding phase were tagged by the incidental presentation of both neutral and negative pictures, which appeared one at time. Thus, the effect of emotion on VSWM was investigated as a within-subject factor. In Experiment 1C, object positions during encoding phase were tagged by the incidental presentation of both neutral and positive pictures, which appeared one at time.
These results confirm the hypothesis that emotional information enhances spatial memory performance when emotional and non-emotional stimuli compete with one another for accessing the working memory system competition effect. Moreover, the emotional content of images increased long-term memory for the incidentally learned pictures. Since negative and positive pictures were more activating than neutral ones, a possible effect of emotional arousal can be envisaged which allows emotional pictures to get a priority access to the memory system Mather and Sutherland, However, since negative and positive pictures are also emotionally valenced, in comparison to neutral ones, a possible role of valence in determining a priority access to the memory system cannot be ruled out.
Experiment 2 was specifically aimed at investigating the role of emotional arousal in the competition mechanism. The ABC theory predicts that the level of arousal of emotional stimuli regulates the access to the working memory system, enhancing the encoding of within-object characteristics, like spatial position Mather and Sutherland, Therefore, we expected that increasing the arousal level of negative pictures would lead to an enhancement of spatial working memory performance. In the high-arousal group, rectangles were tagged by the presentation of neutral and high-arousal negative pictures while in the low-arousal group, rectangles were tagged by neutral and negative pictures with low arousal values Table 1.
Statistical analysis two-way ANOVA carried out on the displacement errors considering arousal high vs. Figure 3. Effect of arousal High and Low Negative vs. Neutral on VSWM. Mean displacement error pixel displayed by participants in the immediate re-location test. These results indicate that arousal manipulation, obtained by selecting negative pictures with different level of arousal high vs.
Since arousal did not seem to significantly impact on the competition between negative and neutral information for accessing the working memory system, we explored the possible effect of valence in this competition. Thus, in Experiment 3A we tagged the object position with negative and neutral pictures but this time pictures had comparable arousal levels Table 1.
In Experiment 3B, rectangle position was tagged by positive and neutral pictures with comparable arousal values Table 1. In both cases valence is a within-subject factor. Figure 4. Effect of valence Negative or Positive vs. Mean displacement error pixel displayed by participants in the immediate object re-location test A when neutral and negative information with similar level of arousal is presented within the same encoding trial, and B when neutral and positive information with similar level of arousal is presented within the same encoding trial.
The results of these experiments indicate that when arousal is kept constant between neutral and negative pictures, or between positive and neutral pictures, valence significantly affects visuo-spatial performance. Experiment 4 was designed in order to verify the competition hypothesis when neutral stimuli are not presented, and competition within the same trial is only among emotional pictures with different valence positive and negative values.
Thus, object positions were tagged by negative and positive pictures, with comparable arousal levels Table 1. Interestingly, the displacement error of both positive- and negative-related objects Figure 5. Effect of valence Positive vs. Negative on VSWM. Mean displacement error pixel displayed by participants in the immediate object re-location test when negative and positive information with similar level of arousal is presented within the same encoding trial.
Bars: standard error mean. In the delayed free recall test, both positive and negative pictures of this experiment were better remembered 0. Notably, in this case, both positive and negative pictures had a higher arousal compared to the neutral pictures of Experiment 1B. Taken together the results of this experiment indicate that when competition occurs among stimuli that all have an emotional value either positive or negative , the effect of emotion on working memory performance vanishes.
The present study was aimed at investigating the effect of the incidental presentation of emotional stimuli on VSWM. To this purpose an object-relocation task was used in which emotional stimuli appeared superimposed on the objects to be relocated.
The first hypothesis tested was that the emotional content of stimuli affects memory for object position only when neutral and emotional stimuli are in competition with one another, i.
Previous studies on the effect of emotion on spatial working memory have yielded contrasting results: positive, negative or no effects of processing stimuli with an emotional content on spatial working memory performance have been reported Putman et al. For example, there is evidence showing that emotions have a negative impact on spatial working memory because the emotional content of stimuli captures attention, subtracting resources for processing other within-object features, like spatial location Mather et al.
The possibility that the attentional bias exerted by emotional stimuli distracts from the main task, thus impairing memory performance, has been suggested Dolcos and McCarthy, ; Bannerman et al. Although emotional stimuli are able to capture attention, there are studies showing no effect or even an improving effect of emotion on spatial working memory performance Bannerman et al. Bannerman et al. The authors explained the lack of effect by considering the possibility that the attentional bias exerted by the emotional content of stimuli is specifically directed to object-identity recognition i.
Gonzalez-Garrido et al. The authors suggested that emotional faces attracted more attention increasing spatial memory performance through a domain-general attention-based mechanism Gonzalez-Garrido et al. Instead, if stimuli encountered all had a similar emotional impact like in our Experiment 1A , the enhancing effect on spatial working memory performance, observed in the competition condition, disappears. The enhancing effect of emotion on spatial working memory performance in competition condition has been replicated also with neutral and positive stimuli Experiment 1C.
The advantage yielded by arousing stimuli is maintained in the working memory system, where emotionally tagged information dominates the competition for mental resources Richter-Levin and Akirav, ; Mather and Sutherland, ; Lee et al.
In the absence of competition, when all stimuli have a similar emotionally arousing content, like in Experiment 1A, it is plausible to hypothesize that the emotional content of stimuli captures the attentional resources, facilitating the encoding of within-object features. However, when competition between the mental representation of emotional and non-emotional stimuli is lacking due to the fact that all the stimuli have a similar emotional impact , no prioritization effect emerges during stimulus processing in working memory.
This could explain why emotion did not affect spatial working memory performance in Experiment 1A. To the best of our knowledge, this is the first time that the competition hypothesis is evaluated by comparing the results of two experimental manipulations in which the emotional stimuli compete or not for accessing to working memory. The results of our experiments indicate that competition among stimuli is an important factor to consider when the effect of emotion on memory performance is investigated.
In Experiment 2 we sought to verify the prediction that arousal is the emotional dimension that mainly determines the enhancement in spatial working memory performance for the relocation of negative-related objects. A better location memory has been found for arousing pictures independently of their valence, suggesting that arousal, rather than valence, is the critical dimension for the emotion-enhancing memory effect Mather and Nesmith, Therefore, we hypothesized that increasing the arousal level of negative pictures should lead to an improvement in spatial memory performance.
Overlapping performances in relocating objects associated with high- and low-arousing pictures were observed Experiment 2 , although both high and low arousing negative-related objects were better relocated than neutral-related objects. Even if we cannot definitely rule out the possibility that this pattern of results is linked to a floor effect, our findings do not seem to be in line with the prediction that arousal is the main emotional dimension affecting VSWM.
On the other hand, like for previous experiments, we observed a better memory performance for negative pictures than for neutral ones in the delayed 24 h free recall task.
0コメント