Psychonomic Bulletin & Review 2010, 17 (5), 680-686 doi:10.3758/PBR.17.5.680
Improving visual short-term memory by sequencing the stimulus array NIKLAS IHSSEN, DAVID E. J. LINDEN, AND KIMRON L. SHAPIRO Bangor University, Bangor, Wales When multiple objects are presented briefly and simultaneously in a visual array, visual short-term memory (VSTM) can maintain only a limited number of these items. The present research report reveals that splitting the to-be-remembered items into two sequential arrays significantly increases VSTM performance relative to the simultaneous presentation of the same items. A memory benefit also emerges when the full object array is flashed twice (repeated) rather than being presented continuously for the same duration. Moreover, the sequential and repetition benefits are specifically pronounced for individuals with low performance for simultaneously presented items. Our results suggest that the conventional, simultaneous presentation mode may underestimate VSTM performance due to attentional limitations and/or competition between stimulus representations. In contrast, temporal segregation of the stimulus input may help participants maximize their performance and utilize their full VSTM capacity.
To estimate the amount of information retained in visual short-term memory (VSTM), the traditional changedetection method (cf. Pashler, 1988) presents a shortduration memory array containing a varying number of objects drawn from a finite stimulus set (e.g., colored squares), which are to be retained across a brief memory interval. The memory array is followed by a test array in which one object is different on 50% of trials and participants are asked to report whether the test array is the same as the memory array previously presented on that trial. Using this approach, VSTM has been investigated across a variety of experimental manipulations, such as testing for whole objects or features, short versus long encoding time, and different response requirements (Luck & Vogel, 1997). Although VSTM capacity varies across individuals (typically 2–4 items; Cowan, 2001), it appears to be relatively fixed within an individual. Recent findings suggest that interindividual capacity differences may originate from differences in attentional control (Vogel, McCollough, & Machizawa, 2005). High-capacity participants efficiently select relevant items from a stimulus array for storage. In contrast, low-capacity participants show deficient selective attention leading to a lower observed capacity, even though absolute storage capacity may be the same. The present studies examine whether sequencing the stimulus array may affect attentional control, in turn increasing VSTM performance1—that is, the amount of information that an individual is able to extract and/or retain in the change-detection task. Previous attempts to find a method by which VSTM can be improved have been creative and varied. Olson and colleagues (Olson & Jiang, 2004; Olson, Jiang, & Moore,
2005) sought to improve change-detection performance by training. Whereas repetition of a given memory display led to spatial associations—that is, facilitated detection when the same locations were probed across display repetitions—it did not improve VSTM per se. There have also been several unsuccessful attempts to improve VSTM by temporally segregating the to-be-remembered stimuli. Using a method prescient of our study, Kumar and Jiang (2005) presented two successive memory arrays at varying SOAs (between 27 and 1,277 msec). VSTM performance did not vary as a function of the temporal gap between the two arrays, regardless of the nature of the changedetection task (e.g., spatial locations, colors, and orientations). However, Vogel, Woodman, and Luck (2006) determined that it takes approximately 50 msec per item to successfully encode and consolidate objects into VSTM. Because Kumar and Jiang presented each memory array for 27 msec, these investigators may not have provided sufficient time for encoding and/or consolidation. On the other hand, they did not measure the amount of information stored in VSTM for simultaneously presented objects and therefore did not allow conclusions about performance changes relative to “baseline.” Although it was not the purpose of their study to do so, Xu and Chun (2006) failed to find VSTM performance differences between trials in which objects were presented simultaneously for 200 msec and trials in which each object appeared successively, with each object displayed for 50 msec. In another experiment, no differences emerged when four objects were displayed either simultaneously or two at a time for an identical display duration. A similar outcome was obtained by Johnson, Spencer, Luck, and Schöner (2009),
N. Ihssen,
[email protected]
© 2010 The Psychonomic Society, Inc.
680
IMPROVING VSTM
681
who displayed three objects simultaneously for 500 msec or successively for 200 msec each. Although the above experiments employed sequential displays, as were used in the present experiment, important procedural differences such as the number and/or complexity of items displayed at any one time likely account for the different outcomes observed.
350 msec). This resulted in a halving of single objects’ encoding time in the simultaneous condition relative to the sequential condition. In the present experiments, the single memory display in the simultaneous condition was presented for twice the length of time (700 msec) as were each of the two displays in the sequential condition (350 msec each).
PARADIGM
EXPERIMENT 1
We examined VSTM performance using a variant of the change-detection task similar to that reported by Fecteau and Shapiro (2008). In the present experiments, participants were asked to remember two briefly presented displays showing, in each trial, a fixed number of geometric objects (four colored squares and four white shapes randomly drawn from a pool of eight objects in each category). The amount of information retained in VSTM was gauged using a subsequently presented test display that showed, in an equal proportion of experimental trials, the set from either the colored squares or white shapes. Critically, participants did not know which of the two sets would be tested. They indicated by buttonpress whether the four objects in the test display were the same (the “s” key) or different (the “d” key), compared with the previously shown colors/shapes. The two four-object sets (colors and shapes) were presented either in two temporally separated displays (the sequential condition) or in a single display of eight objects (the simultaneous condition). The latter condition was used as an individual baseline measure, assessing VSTM performance in the canonical way. In a third condition (the repetition condition), the single eight-object display was presented twice. All conditions were matched for display time. The third condition was introduced to test the notion that presenting the memory array twice, regardless of content, enhances memory performance. It is conceivable that rather than splitting/sequencing the eight objects into two halves (making it a “two-page display”), any stimulus that triggers two temporal encoding/retention episodes may improve VSTM. The rationale for using two stimulus categories (colors and shapes), rather than drawing objects from one category, was to provide equal opportunity to group objects across conditions. It can be argued that a benefit in the sequential condition might simply reflect easier spatial grouping for two temporally separated displays. By categorically distinguishing the two memory sets, grouping strategies could in principle be applied in the simultaneous condition, as well. Importantly, in both experiments, colored squares and white shapes appeared in nonoverlapping regions of the stimulus array. Previous results (Fecteau & Shapiro, 2008) have indicated that spatial separation of the two memory sets is crucial to prevent visual masking, which may lead to an underestimation of memory performance in the sequential condition. One potential confound in the report by Fecteau and Shapiro (2008) concerns the time that each individual object was available for encoding; in their experiments, all displays were presented for the same duration (i.e.,
Method Participants. Twenty-four volunteers from Bangor University (15 women, M age 24.0 years, SD 7.9) with normal or corrected-to-normal vision gave informed consent to participate in Experiment 1. For their participation, they received course credits. The School of Psychology research ethics committee approved both experiments. Stimuli and Procedure. Stimuli were presented against a gray background on a 19-in. 100-Hz monitor. At a viewing distance of 70 cm, single objects subtended a visual angle of 1.5º 1.5º (full array, 3.9º 9.0º). Participants were instructed to maintain fixation on a gray cross (0.3º 0.3º) that remained at the center of the screen throughout each trial. Colors and shapes were presented in groups either above or below fixation. Assignment to the upper or lower visual field was counterbalanced across participants. Within each condition, half of the trials probed either colored squares or white shapes and either a “same” or “different” display. For “different” test displays, we replaced one of the objects shown in the previous displays by another (new) object from the same set and in the same location. As is illustrated in Figure 1, stimuli in each trial were preceded by a fixation baseline of 500 msec. In the sequential and repetition conditions, the two memory displays were presented for 350 msec each and followed by a retention interval of 950 msec showing a blank screen and the fixation cross. In the sequential condition, the first display showed colors in half of the trials and shapes in the other half, in a random order. In the simultaneous condition, a single memory display was shown for 700 msec. The onset of the single display was matched in an equal proportion of the trials to the onset of either the first or second display in the sequential condition. The test display remained on the screen until participants pressed either the “s” or the “d” key. After responding, participants received visual feedback about the correctness of their response and then were prompted to initiate the next trial. The three conditions were presented in separate blocks of 160 trials each, with block order balanced across participants. Before each block, participants completed a practice phase. To prevent verbal encoding strategies, participants performed an articulatory suppression task. Here, a low-frequency three-syllable word (e.g., “peppermint”) was presented before each of the three blocks, which participants had to repeat silently throughout the memory trials. Data analysis. We estimated VSTM performance using the formula proposed by Cowan (2001): k (hit rate correct rejection rate 1) * set size. Using a set size of four, we calculated separate k values for trials in which colors versus shapes were probed. As a result, each k value represented the number of items retained from the probed set of four items. We then added these two values to create a single measure of change-detection performance. For each participant, k values were computed for each of the three conditions and submitted to a repeated measures ANOVA using presentation mode (three: simultaneous, sequential, repetition) as a withinsubjects factor. Significant effects were followed by paired t tests. To reveal effects of temporal position, we also computed k values for specific subconditions. In the sequential condition, we differentiated trials that probed the first versus second display. In the simultaneous condition, we calculated separate k values for displays whose onset time corresponded either to the first or the second sequential display. We entered these k values in a combined ANOVA using presentation
682
IHSSEN, LINDEN, AND SHAPIRO
A
1st Memory Display
2nd Memory Display
Test Display
Simultaneous (Baseline)
+
Sequential
+
+
+
Repetition
+
+
+
+
+
t
B
1st Display
2nd Display
Test Display
Sim 1st
8
0
4
Sim 2nd
0
Seq 1st
4
4
4 (Test 1st)
Seq 2nd
4
4
4 (Test 2nd)
Rep
8
8
500
350
8
950
4
350
4
950
Until Response
t (msec)
Figure 1. (A) Examples of the stimuli used in Experiment 1. (B) Trial schematics showing the stimulus timing for the three conditions in both experiments (Sim, simultaneous displays shown at 1st vs. 2nd position; Seq, sequential 1st vs. 2nd displays tested; Rep, repetition). Dark bars illustrate the duration of stimulus displays and include the numbers of items presented; gray bars represent the duration of fixation intervals.
mode (two: simultaneous, sequential) and temporal position (two: 1st position, 2nd position) as within-subjects factors.
Results and Discussion Presentation mode significantly affected memory performance [F(2,46) 3.91, p .05, h p2 .15]. Follow-up analyses revealed an improvement of VSTM performance for both the sequential condition [t(23) 2.36, p .05] and the repetition condition [t(23) 2.65, p .05], relative to the simultaneous condition (see Figure 2, left panel). This finding is important in two ways: First, a benefit (cf. Fecteau & Shapiro, 2008) for sequential displays is replicable even when the simultaneous display is equated for encoding time by presenting it for twice the duration of the two half-displays in the sequential condition. Second, participants not only remembered more of the eight items when the objects were split into two consecutive groups of four objects but also when the continuous 700-msec dis-
play showing all objects at once was temporally separated into two 350-msec displays (the repetition condition). Closer analysis of the sequential and simultaneous conditions alone showed a significant effect of presentation mode [F(1,23) 5.55, p .05, h 2p .19] and temporal position [F(1,23) 67.15, p .001, h 2p .75], with the latter reflecting a recency benefit, which is a welldocumented phenomenon in memory tasks (Broadbent & Broadbent, 1981). We also found an interaction between presentation mode and temporal position [F(1,23) 13.30, p .01, h 2p .37]. Comparisons between sequential and simultaneous displays that corresponded to each other with regard to onset time of the probed display showed the following: Performance was specifically improved for sequential displays when the second array was probed (see Figure 2, right panel) [t(1,23) 4.86, p .001], but not when the first array was probed [t(1,23) 0.42].
IMPROVING VSTM
3
2
2
1
Sim
Seq
Rep
1
Sim
Seq
Figure 2. (A) Estimates of visual short-term memory (VSTM) performance using Cowan’s k for simultaneous (Sim), sequential (Seq), and repeated (Rep) displays in Experiment 1. (B) Effects of temporal position and presentation mode on k scores in the simultaneous and sequential conditions of Experiment 1. Note that “1st”/“2nd” in the sequential condition refers to which of the two four-item displays was tested and, in the simultaneous condition, to the temporal position at which the eight-item display appeared. Error bars indicate the standard error of the mean.
Crucially, however, when k values for first and second displays were summed, the sequential condition still yielded higher performance estimates than did the simultaneous condition [t(23) 2.36, p .05]. This result indicates that participants did not use a trade-off strategy leading to better performance for the second display at the expense of the first display. Moreover, a comparison between k values for colors and shapes showed that overall performance was better in color trials [F(1,23) 33.16, p .001, h 2p .59]. Importantly, this effect did not interact with presentation mode [main effect, F(2,46) 3.9, p .05, h 2p .15; interaction, F(2,46) 1.0], indicating that participants did not use configural or relational shape information more frequently or efficiently in the sequential or repetition, in comparison with the simultaneous condition. Interestingly, we found a relationship between individuals’ baseline performance (average k value in the simultaneous conditions) and the magnitude of their performance increase (see Figure 3). The difference in k between the sequential 2nd position and simultaneous 2nd position conditions correlated negatively with baseline performance (r .45, p .05). The same was true for the relationship between the benefit in the repetition condition and baseline performance (r .46, p .05). Thus, the lower an individual’s VSTM performance, the larger is the gain from sequencing the to-be-remembered objects/displays. EXPERIMENT 2 The main goal of Experiment 2 was to control for perceptual load in the stimulus displays. Balanced perceptual displays allowed us to rule out a potential confound of Experiment 1—namely, that the sequential benefit was due to a reduction in performance in the simultaneous condition as
A Sequential Benefit (k Difference)
3
a result of visual crowding in the eight-item as opposed to the four-item displays (Strasburger, Harvey, & Rentschler, 1991). In Experiment 2, we embedded all memory objects in a bilateral symmetrical array of placeholders (the letter “H”), matched in vertical size to the geometric objects. By this means, the overall number of objects present on the screen at any given point in time was identical across conditions. Furthermore, we inserted an “empty” placeholder display in the simultaneous condition, matched to the start and duration of the first/second sequential display. The results from Experiment 1 suggest that task strategies—for example, attending only to the second display or encoding configural information—did not play a major role in the observed sequential and repetition benefits. Nevertheless, Experiment 2 further counteracted the use of strategies by presenting the three conditions randomly rather than blocked.
3.5 3 2.5 2 1.5 1 0.5 0
0.5 1 1.5 0
1
2
3
4
5
6
5
6
Baseline Performance (k) 2
Repetition Benefit (k Difference)
4
2nd
4
B
1st
5
2nd
A
1st
VSTM Performance (k)
5
683
B
1.5 1 0.5 0
0.5 1
1.5 0
1
2
3
4
Baseline Performance (k) Figure 3. (A) Relationship between benefit scores (k for 2nd sequential displays k for 2nd simultaneous displays) of single participants and their baseline performance (average k across 1st and 2nd simultaneous displays). Triangles above the dotted line (0) indicate a performance increase, whereas triangles below the dotted line show a performance decrease. The solid regression line illustrates a negative relation between individuals’ benefits and their baseline performance. (B) Relationship between benefit scores for repeated displays (k repetition condition mean k for simultaneous displays) of single participants and their mean baseline performance.
684
IHSSEN, LINDEN, AND SHAPIRO 1st Memory Display
Simultaneous (Baseline)
Sequential
H
H
H H H H H H
2nd Memory Display
+
H H H H H H H H H H
H H H H H H H H H H
+
H H H H H H H H H H
H H H H H H
+
H H H H H H H H H H
+
H H H H H H H H H H
t Figure 4. Examples of simultaneous and sequential displays used in Experiment 2. Placeholders and an “empty” array in the simultaneous condition were included to equate the conditions for visual information and perceptual load.
4
4
3
3
2
2
1
B
1
Sim
Seq
Rep
Sim
2nd
5
1st
A 5
2nd
Results and Discussion The overall k level was substantially lower than in the previous study, indicating higher task difficulty in Experiment 2, likely due to increased selection demands. Crucially, however, using a placeholder array to control for perceptual effects did not change the pattern of results: ANOVA showed a main effect of presentation mode [F(2,40) 9.73, p .001, h 2p .33], reflecting increased performance in the sequential [t(20) 4.21, p .001] and in the repetition conditions [t(20) 4.36, p .001], relative to the simultaneous condition (see Figure 5, left panel). Accordingly, a cross-experimental ANOVA using Experiment (1 vs. 2) as an additional between-subjects factor revealed a main effect of experiment [F(1,43) 16.69, p .001, h 2p .28] and presentation mode [F(2,86) 13.23, p .001, h 2p .24], but no interaction between these factors [F(2,86) 1.58, p .21]. Moreover, when conducting an ANOVA on the k scores in the sequential and repetition subconditions, we found
main effects of temporal position [F(1,20) 19.44, p .001, h 2p .49] and presentation mode [F(1,20) 17.76, p .001, h 2p .47], but no interaction [F(1,20) 1.52, p .23]. Performance significantly increased for sequential displays at the 2nd position relative to corresponding simultaneous displays [t(20) 5.16, p .0001]. There was also a trend toward higher k values for the comparison between sequential and simultaneous displays at the 1st position [t(20) 1.77, p .091], in contrast to Experiment 1. Again, we found a relationship between the degree of the sequential benefit and participants’ baseline k level. The significant negative correlation between mean performance for simultaneous displays and k increase for 1st
1st
Participants. Twenty-one volunteers (11 women, M age 22.57 years, SD 6.35) consented to participate in Experiment 2. Participants had normal or corrected-to-normal vision and received course credits for participation. Stimuli and Procedure. Apart from the changes noted below, Experiment 2 was conducted identically to Experiment 1. The two memory sets were presented together at either the left or right side of fixation2 and embedded within a bilateral array of 2 10 placeholders (see Figure 4). Colored squares covered randomly 4 of 5 placeholders above fixation, whereas white shapes appeared at 4 of the 5 placeholders below fixation. Assignment of colors/shapes to the upper or lower quadrant was counterbalanced across participants. Trial timing was identical to Experiment 1, except for an “empty” placeholder array now occurring in the simultaneous condition at temporal positions matched to the sequential condition. In contrast to the previous study, conditions were randomized.
VSTM Performance (k)
Method
Seq
Figure 5. (A) Cowan’s k for simultaneous (Sim), sequential (Seq), and repeated (Rep) displays in Experiment 2. (B) The k scores for 1st and 2nd displays in the simultaneous and sequential conditions of Experiment 2. VSTM, visual short-term memory.
IMPROVING VSTM sequential versus 1st simultaneous displays (r .47, p .05), indicated stronger benefits for low-performing participants. For the repetition benefit, there was a trend toward this pattern (r .41, p .069). GENERAL DISCUSSION The present experiments are important in revealing a significant increase in VSTM performance when the tobe-retained items appear in two temporal frames, rather than in the canonical single temporal frame. This increase occurs even though the duration of the simultaneous display is matched to the total duration of the two displays in the repetition and sequential conditions. Importantly, we showed that perceptual factors such as crowding or increased perceptual load in the simultaneous condition do not contribute to the present effect. The intriguing question to consider is what mechanism might underlie the present pattern of results. It has been argued that there exists a fragile form of VSTM (Sligte, Scholte, & Lamme, 2008) in which up to 16 objects can be stored for approximately 1,000 msec after stimulus offset. The fragile, high-capacity store is assumed to intervene between iconic memory and a more stable, capacitylimited form of VSTM. This high capacity, however, has only been revealed using “retro cues” in the changedetection task. The retro-cue method inserts spatial cues during the retention interval before the probe display occurs, and indicates the location of the item that is subject to a possible change. Because the retro-cue paradigm was not used in the present paradigm, fragile VSTM is most likely not a candidate to account for the increased performance witnessed in the present studies. We considered the possibility that the observed performance does not represent an increase in the number of STM representations but, alternatively, the involvement of long-term memory (LTM). One could argue that by the time the second display arrived in the sequential condition, the first display had been transferred into LTM. However, given the present stimulus durations and stimulus–test intervals—and the articulatory suppression requirement—we think it unlikely that participants were able to establish LTM traces. A more likely account for the present findings is that splitting or repeating the stimulus input overcomes the limitations of processing resources that arise when multiple items have to be encoded and consolidated at the same time (Mayer et al., 2007). As indicated by the work of Vogel et al. (2005), observed VSTM capacity is closely linked to the ability to attend to relevant items in the stimulus array and to exclude irrelevant items (distractors) from being stored, respectively. Although in our paradigm, participants did not have to select relevant from irrelevant items, facilitated selective attention may underlie the presently reported VSTM improvement as well. Specifically, allocating attention to one half of a stimulus array first (e.g., colors), and then to the other half (shapes) later on, may be more efficient than attending to both halves at once. In the repetition condition, mimicking the sequential condition,
685
the occurrence of two temporal frames may have facilitated the selection of two different object subsets across the two displays. The above ideas are also supported by our finding of a negative correlation between baseline performance and k increase in the sequential/repetition conditions. Consistent with Vogel et al. (2005), sequencing the stimulus array was specifically beneficial for lowperforming participants—that is, individuals with suboptimal selective attention processes. Similarly, Fukuda and Vogel (2009) showed that susceptibility to attentional interference exerted by irrelevant distractors is closely linked to individual differences in VSTM capacity, which provides an account for the reduced performance estimate observed in Experiment 2 relative to Experiment 1, where the former used placeholders requiring suppression. A related account for the present findings is suggested by biased competition models (Desimone & Duncan, 1995), whose mechanisms operate on simultaneously presented stimuli. According to biased competition, the brain responds to simultaneous stimuli by selecting a small subset—possibly one—to be attended, in turn causing less activation, even suppression, of the others. In the sequential condition, the amount of biased competition may have been reduced because the display elements were split into consecutive halves. In the repetition condition, competition may have been reduced because participants attended to two different object subsets across the two displays. AUTHOR NOTE The present study was supported by grants to all the authors from the Wales Institute of Cognitive Neuroscience. Correspondence concerning this article should be addressed to N. Ihssen, Wales Institute of Cognitive Neuroscience/Wolfson Centre for Clinical and Cognitive Neuroscience, School of Psychology, Bangor University, Gwynedd LL57 2AS, Wales (e-mail:
[email protected]). REFERENCES Broadbent, D. E., & Broadbent, M. H. P. (1981). Recency effects in visual memory. Quarterly Journal of Experimental Psychology, 33A, 1-15. Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral & Brain Sciences, 24, 87-114; discussion, 114-185. doi:10.1017/S0140525X01003922 Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193-222. doi:10.1146/annurev.ne.18.030195.001205 Fecteau, J., & Shapiro, K. (2008). Multiplying the capacity of visual working memory [Abstract]. Journal of Vision, 8(6), 1168a. Fukuda, K., & Vogel, E. K. (2009). Human variation in overriding attentional capture. Journal of Neuroscience, 29, 8726-8733. doi:10.1523/JNEUROSCI.2145-09.2009 Johnson, J. S., Spencer, J. P., Luck, S. J., & Schöner, G. (2009). A dynamic neural field model of visual working memory and change detection. Psychological Science, 20, 568-577. doi:10.1111/j.1467 -9280.2009.02329.x Kumar, A., & Jiang, Y. (2005). Visual short-term memory for sequential arrays. Memory & Cognition, 33, 488-498. Luck, S. J., & Vogel, E. K. (1997). The capacity of visual working memory for features and conjunctions. Nature, 390, 279-281. doi:10.1038/36846 Mayer, J. S., Bittner, R. A., Nikolic, D., Bledowski, C., Goebel, R., & Linden, D. E. (2007). Common neural substrates for visual working memory and attention. NeuroImage, 36, 441-453. doi:10.1016/ j.neuroimage.2007.03.007
686
IHSSEN, LINDEN, AND SHAPIRO
Olson, I. R., & Jiang, Y. (2004). Visual short-term memory is not improved by training. Memory & Cognition, 32, 1326-1332. Olson, I. R., Jiang, Y., & Moore, K. S. (2005). Associative learning improves visual working memory performance. Journal of Experimental Psychology: Human Perception & Performance, 31, 889-900. doi:10.1037/0096-1523.31.5.889 Pashler, H. (1988). Familiarity and visual change detection. Perception & Psychophysics, 44, 369-378. Sligte, I. G., Scholte, H. S., & Lamme, V. A. F. (2008). Are there multiple visual short-term memory stores? PLoS ONE, 3, e1699. doi:10.1371/journal.pone.0001699 Strasburger, H., Harvey, L. O., Jr., & Rentschler, I. (1991). Contrast thresholds for identification of numeric characters in direct and eccentric view. Perception & Psychophysics, 49, 495-508. Vogel, E. K., & Machizawa, M. G. (2004). Neural activity predicts individual differences in visual working memory capacity. Nature, 428, 748-751. doi:10.1038/nature02447 Vogel, E. K., McCollough, A. W., & Machizawa, M. G. (2005). Neural measures reveal individual differences in controlling access to working memory. Nature, 438, 500-503. doi:10.1038/nature04171
Vogel, E. K., Woodman, G. F., & Luck, S. J. (2006). The time course of consolidation in visual working memory. Journal of Experimental Psychology: Human Perception & Performance, 32, 1436-1451. doi:10.1037/0096-1523.32.6.1436 Xu, Y., & Chun, M. M. (2006). Dissociable neural mechanisms supporting visual short-term memory for objects. Nature, 440, 91-95. doi:10.1038/nature04262 NOTES 1. It should be noted that we prefer the term VSTM performance over VSTM capacity because we do not want to imply that the present paradigm is capable of altering storage “capacity,” per se, or any hypothetical upper capacity limit. 2. A lateralized stimulus configuration was used to afford later recording of the contralateral delay activity, which is considered an electrophysiological marker of VSTM load (Vogel & Machizawa, 2004). (Manuscript received December 23, 2009; revision accepted for publication March 27, 2010.)