Developmental change in working memory strategies: from passive maintenance to active refreshing

Change in strategies is often mentioned as a source of memory development. However, though performance in working memory tasks steadily improves during childhood, theories differ in linking this development to strategy changes. Whereas some theories, such as the time-based resource-sharing model, invoke the age-related increase in use and efficiency of a strategy of active maintenance of memory traces, other theories, such as the task-switching model, do not mention strategy change. According to these models, either the cognitive load of the task or the duration of maintenance would account for recall performance. In the present study, we varied orthogonally these 2 factors. The results revealed that a different and unique factor affected recall performance at different ages: the duration of maintenance at age 6 and the cognitive load at age 7. As described by the task-switching model, younger children would not implement any maintenance activities while performing a concurrent task, their memory traces suffering from a time-based decay. This suggests that an increasing capacity of cognitive monitoring allows children to shift from this passive maintenance of memory traces to the active refreshing thereof at around the age of 7, reunifying the 2 current accounts of working memory development as 2 developmental stages.     

Time constraints and resource sharing in adults' working memory spans

This article presents a new model that accounts for working memory spans in adults, the time-based resource-sharing model. The model assumes that both components (i.e., processing and maintenance) of the main working memory tasks require attention and that memory traces decay as soon as attention is switched away. Because memory retrievals are constrained by a central bottleneck and thus totally capture attention, it was predicted that the maintenance of the items to be recalled depends on both the number of memory retrievals required by the intervening treatment and the time allowed to perform them. This number of retrievals:time ratio determines the cognitive load of the processing component. The authors show in 7 experiments that working memory spans vary as a function of this cognitive load.     

On the law relating processing to storage in working memory

Working memory is usually defined in cognitive psychology as a system devoted to the simultaneous processing and maintenance of information. However, although many models of working memory have been put forward during the last decades, they often leave underspecified the dynamic interplay between processing and storage. Moreover, the account of their interaction proposed by the most popular A. D. Baddeley and G. Hitch's (1974) multiple-component model is contradicted by facts, leaving unresolved one of the main issues of cognitive functioning. In this article, the author derive from the time-based resource-sharing model of working memory a mathematical function relating the cognitive load involved by concurrent processing to the amount of information that can be simultaneously maintained active in working memory. A meta-analysis from several experiments testing the effects of processing on storage corroborates the parameters of the predicted function, suggesting that it properly reflects the law relating the 2 functions of working memory.     

Time-related decay or interference-based forgetting in working memory?

The time-based resource-sharing model of working memory assumes that memory traces suffer from a time-related decay when attention is occupied by concurrent activities. Using complex continuous span tasks in which temporal parameters are carefully controlled, P. Barrouillet, S. Bernardin, S. Portrat, E. Vergauwe, & V. Camos (2007) recently provided evidence that any increase in time of the processing component of these tasks results in lower recall performance. However, K. Oberauer and R. Kliegl (2006) pointed out that, in this paradigm, increased processing times are accompanied by a corollary decrease of the remaining time during which attention is available to refresh memory traces. As a consequence, the main determinant of recall performance in complex span tasks would not be the duration of attentional capture inducing time-related decay, as Barrouillet et al. (2007) claimed, but the time available to repair memory traces, and thus would be compatible with an interference account of forgetting. The authors demonstrate here that even when the time available to refresh memory traces is kept constant, increasing the processing time still results in poorer recall, confirming that time-related decay is the source of forgetting within working memory.     

Stroke damages attentional maintenance in working memory

Stroke is the main cause of acquired disability in adults, and specific deficits in working memory (WM) are among the most common cognitive consequences. In neuropsychological routine, WM is most of the time investigated in the framework of the multicomponent model (Baddeley & Hitch, 1974, The psychology of learning and motivation, 47). Using a more recent theoretical WM model, the time-based resource-sharing (TBRS) model (Barrouillet et al., 2011, Psychol. Rev., 118, 175), the aim of the present study was to investigate in young post-stroke patients to which extent attentional maintenance is impaired in WM. To address this question, we discarded other factors known to directly influence WM performance, that is processing speed and short-term memory span. We proposed to 53 post-stroke patients and to 63 healthy controls a complex span paradigm in which participants were asked to alternate between the memorization of a series of images and a concurrent parity judgement task of a series of digits. To investigate the attentional maintenance processes, we manipulated the cognitive load (CL) of the concurrent task. CL effect is typically interpreted as the involvement of attentional maintenance processes. The task was adapted to each participant according to their processing speed and memory span. As expected, the results showed higher recall performance in healthy controls compared with post-stroke patients. Consistent with the literature, we also observed higher performance when the CL was low compared with high. However, the improvement in recall at low CL was smaller for post-stroke patients compared with controls, suggesting that post-stroke WM deficit could be in part due to a deficit of the attentional maintenance processes.     

The effects of processing time and processing rate on forgetting in working memory: Testing four models of the complex span paradigm

Four models of working memory processes in the complex span paradigm were tested: The task-switching model of Towse, Hitch, and Hutton (1998), the interference account of Saito and Miyake (2004), and two versions of the time-based resource-sharing model of Barrouillet, Bernardin, and Camos (2004). On the basis of a reading span paradigm that used segmented sentences, the effect of processing time on the recall of words was investigated while the amount of processing was held constant. Two conditions of reading (continuous vs. normal) were compared in order to study the influence of brief pauses during reading that could be used for articulatory rehearsal. The results favor a version of the time-based resource-sharing model: A faster reading rate had a negative effect on recall. The effect of reading rate was obtained with continuous as well as normal reading, revealing that even continuous articulation does not prevent simultaneous refreshing of memory traces. A second experiment showed that continuous reading made concurrent articulatory rehearsal virtually impossible. These findings imply that a second rehearsal mechanism for verbal working memory, other than articulatory rehearsal, exists.     

Time-related processes in time-based prospective memory and in time-interval production

Time-based prospective memory tasks require the retrieval of previously formed plans (e.g., call me on my mobile phone) either at a specified time (e.g., at 3 p.m.) or after a specified period of time has elapsed (e.g., in 3 min). In the present study, we investigated whether the same time-related processes are recruited for a short-duration time-based prospective memory task and for a time-interval production task. In a dual-task paradigm, we required young adult participants to make category-membership decisions, while they were simultaneously engaged either in a time-based prospective memory tasks or in a time-interval production task. Under one set of conditions, participants were instructed to allocate attention equally to the two tasks, whereas in another set of conditions, they allocated attention primarily to the time-relevant task. The findings revealed that time estimates were not affected by the attention allocation manipulation, although it influenced performance on the concurrent category membership decision making task, and as expected time estimates were more accurate on the prospective memory task than on the time-production task. Based on the preponderance of the findings, we suggest that timing is managed by different mechanisms and/or strategies on time-interval production tasks and on short-duration time-based prospective memory tasks.     

Working memory in children: a time-constrained functioning similar to adults

Within the time-based resource-sharing (TBRS) model, we tested a new conception of the relationships between processing and storage in which the core mechanisms of working memory (WM) are time constrained. However, our previous studies were restricted to adults. The current study aimed at demonstrating that these mechanisms are present and functional before adulthood. For this purpose, we investigated the effect on maintenance of the duration of the attentional capture induced by processing. In two experiments using computer-paced WM span tasks, 10-year-olds were asked to maintain letters while performing spatial location judgments. The duration of this processing was manipulated by varying either the discriminability between target locations or the contrast between targets and background. In both experiments, longer processing times resulted in poorer recall, as we observed previously in adults. These findings suggest that the core mechanisms of WM described by the TBRS model are already settled during childhood.     

On the interpretation of working memory span in adults

Experimental research into children's working memory span has shown that retention duration contributes substantially to span performance, while processing efficiency need not be related to concurrent memory load (Towse, Hitch, & Hutton, 1998). These findings have been used to argue for a model of working memory span that emphasizes time-based forgetting rather than the popular resource-sharing or tradeoff framework. The present paper considers whether adults perform working memory span tasks in a qualitatively different way. Data from reading span and operation span tasks show that adults' performance can be distinguished from that of children, but also that a task-switching model of working memory span can explain some important aspects of performance.     

Do two separate speech measures constrain verbal short-term memory in children?

Previous research has indicated that 2 processing rates may constrain verbal short-term memory performance. These have been linked to individual differences in (a) the time taken to articulate spoken words and (b) the duration of pauses that occur between words in the output responses to memory tasks. Two experiments examined whether evidence for these effects on memory can be obtained for measures taken from a single speech sample. Children articulated pairs of words as rapidly as possible. In both experiments, the spoken duration of words and the length of the pauses between them predicted significant variance in verbal short-term memory performance. It is argued that the duration of words is linked to memory performance through the processes underlying time-based forgetting in short-term memory. In contrast, the duration of pauses in speeded articulation may index individual differences in speech planning processes.     

时距加工的记忆负荷效应

为了探索非时间任务短时记忆加工对不同长度时距加工的影响是否一致,本研究采用双任务范式,操纵时距编码阶段的短时记忆加工负荷,以平均复制时距为因变量指标,考察了非时间任务短时记忆加工对1-6秒时距加工的影响。结果发现,1秒、2秒及3秒的平均复制长度不随记忆负荷的大小而变化,4秒、5秒及6秒的平均复制长度随着记忆负荷变大而缩短。这意味着非时间短时记忆对时距加工影响具有选择性。     

Time estimation and concurrent nontemporal processing: Specific interference from short-term-memory demands

Previous studies have shown that the effect of concurrent nontempcraLprocessmg on time astimation may vary depending on the level of difficulty of the nontemporal task. This is commonly interpreted within the context of so-called distraction/interruption models of temporal processing, which propose that as concurrent task difficulty or complexity is increased, temporal processing receives less attention. We hypothesize that the effect of nontemporal processing does not depend on the level of difficulty as such, but rather on the extent to which the concurrent nontemporal task specifically involves processing in short-term meraery. Four experiments were run in which the short-term memory requirements of concurrent tasks were systematically varied, although all of the tasks were of comparable levels of difficulty. In the first experiment, the effect of memory search on simultaneous temporal productions was proportional to the number of items to search. As with reaction time, produced intervals were shown to increase linearly with the number of items in the memorized set. In Experiment 2, a visual search involving some load on short-term memory interfered in the same way with time production, although to a lesser extent. The last two experiments showed that performing attention-demanding visual search tasks that did not involve short-term memory did not lengthen simultaneously produced time inter-vals. This suggests that interference of nontemporal processing on time processing may not be a matter of nonspecific general purpose attentional resources, but rather of concurrent short-term-memory processing demands.     

工作记忆影响时距加工的初步研究

工作记忆会影响时距加工吗？这是一个值得关注的问题。为此,该研究采用双任务范式,选择1～6秒为考察对象,操纵时距编码阶段的工作记忆负荷,以平均复制时距、比率分数及差异系数为因变量指标,旨在探讨不同工作记忆负荷对时距加工的影响。结果发现,1秒、2秒、3秒平均复制长度、复制准确性及变异性不随记忆负荷的大小而变化,4秒、5秒及6秒的平均复制长度随着记忆负荷变大而缩短、复制准确性随着记忆负荷变大而降低、复制变异性随着记忆负荷变大而增大。这意味着工作记忆对时距加工影响具有选择性,可以由时距加工分层机制假说得以解释。     

Interference between storage and processing in working memory: Feature overwriting,not similarity-based competition

Eight experiments with the complex span paradigm are presented to investigate why concurrent processing disrupts short-term retention. Increasing the pace of the processing task led to worse recall, supporting the hypothesis that the processing task distracts attention from maintenance operations. Neither phonological nor semantic similarity between memory items and processing-task material impaired memory. In contrast, the degree of phonological overlap between memory items and processing-task material affected recall negatively, supporting feature overwriting as one source of interference in the complex span paradigm. When compared directly, phonological overlap impaired memory, but similarity had a beneficial effect. These findings rule out response competition or confusion as a mechanism of interference between storage and processing.     

Insensitivity of visual short-term memory to irrelevant visual information

Several authors have hypothesized that visuo-spatial working memory is functionally analogous to verbal working memory. Irrelevant background speech impairs verbal short-term memory. We investigated whether irrelevant visual information has an analogous effect on visual short-term memory, using a dynamic visual noise (DVN) technique known to disrupt visual imagery (Quinn & McConnell, 1996b). Experiment 1 replicated the effect of DVN on pegword imagery. Experiments 2 and 3 showed no effect of DVN on recall of static matrix patterns, despite a significant effect of a concurrent spatial tapping task. Experiment 4 showed no effect of DVN on encoding or maintenance of arrays of matrix patterns, despite testing memory by a recognition procedure to encourage visual rather than spatial processing. Serial position curves showed a one-item recency effect typical of visual short-term memory. Experiment 5 showed no effect of DVN on short-term recognition of Chinese characters, despite effects of visual similarity and a concurrent colour memory task that confirmed visual processing of the characters. We conclude that irrelevant visual noise does not impair visual short-term memory. Visual working memory may not be functionally analogous to verbal working memory, and different cognitive processes may underlie visual short-term memory and visual imagery.     

Effects of internal clock and memory disorders on duration reproductions and duration productions in patients with Parkinson's disease

Patients with Parkinson's disease (PD) exhibit deficits in perceptual and motor timing as well as impairments in memory and attentional processes that are related to dysfunction of dopaminergic systems in the basal ganglia. The aim of the present study was to assess the relationships existing between impaired duration judgments and defective memory and attention in PD patients. We compared time performance of medicated PD patients and control subjects on a duration reproduction task that is highly memory-dependent, and on a duration production task that could reveal effects of changes in the speed of internal time-keeping mechanisms. Each task was performed in a control counting condition and in a condition requiring divided attention between the temporal task and a concurrent reading task. Moreover, PD patients and control subjects were assessed on memory and attention using a battery of neuropsychological tests. The results revealed that in the concurrent reading condition of the reproduction task, duration judgments tended to be more variable in PD patients than in control subjects. Moreover, variability of duration reproductions was correlated with both measures of memory and of disease severity. In the concurrent reading condition of the production task, duration judgments were significantly shorter in PD patients than in control subjects, and accuracy of duration productions was correlated with scores on sub-tests of short-term memory. The findings suggest that the administration of dopamine did not entirely remove the memory deficits in PD patients. Moreover, DA treatment would have abnormally accelerated the rate of the internal clock leading to shorter duration productions in PD patients. The whole results indicate that dopamine administration in patients might have overshadowed the slowing rate of the internal clock usually reported in non medicated PD patients, without entirely restoring all of the memory functions.     

Visuospatial memory and phonological loop in learning from multimedia

The dual‐task paradigm was used to show how visuospatial working memory and the phonological loop are involved in processing scientific texts and illustrations presented via computer. In experiment 1, two presentation formats were compared: text‐only and text‐with‐illustrations. With a concurrent tapping task, the beneficial effect of illustrations disappeared, while a concurrent articulatory task impaired performance similarly in both presentation formats. An analysis of individual differences revealed that this pattern of results was present in high, but not low spatial span subjects. These results support the selective involvement of visuospatial working memory in processing illustrated texts. In Experiment 2, the text‐only presentation format was compared to an illustrations‐only format. The concurrent articulatory task selectively impaired text‐only processing, compared with processing illustrations‐only. In addition, this pattern of results was found for high, but not low digit span subjects. These results suggest that individual differences define the extent to which the two subsystems of working memory are involved in learning from multimedia. These two subsystems would be mainly involved in the maintenance of a visual trace of illustrations and of a verbatim representation of linguistic information respectively, these representations being the basis for higher‐level comprehension processes. Copyright © 2002 John Wiley & Sons, Ltd.     

The effects of divided attention on encoding processes in memory: mapping the locus of interference.

Despite the detrimental effects of divided attention at encoding on later memory performance, results described in the literature do not unequivocally specify which processes are interrupted during encoding by participants' occupation with a concurrent task. Using a processing analysis framework where the encoding process is viewed as a multiphase mental activity, the current research investigated this issue using a new differential temporal interference paradigm where the study phase of single words was interrupted at different temporal segments. In two experiments, we used performance on both memory and online choice reaction time tasks to assess whether such differential interference would produce different degrees of reduction in participants' later memory performance, as well as changes in the attentional resources required to execute each of the encoding phases. Measures of memory and concurrent task performance in the two experiments converged on similar patterns, showing that all phases of encoding are affected by the concurrent task. However, the initial encoding phase, which is tentatively associated with the initial registration of information, seems especially vulnerable to interference.     

Is the interference between memory processing and timing specific to the use of verbal material?

Increasing load in a memory task performed simultaneously with a timing task shortens perceived time, an effect that has been observed previously with memory tasks using verbal material. The present experiments examine whether two similar memory tasks, one in which verbal material is used and another one in which nonverbal material is used, would produce similar interference effects on concurrent time reproduction. In Experiment 1, the number of nonverbal stimuli (pseudo‐random dot patterns) was manipulated in a memory task performed while a temporal interval to be reproduced was encoded. Reproductions shortened proportionally to the duration of memory processing executed during time estimation. Verbal stimuli (consonants) were used in Experiment 2 in otherwise identical experimental conditions. Effects observed in Experiment 2 were comparable to those obtained in Experiment 1. Taken together, these results support the notion that interference from memory tasks on concurrent time estimation is not determined by the specific type of material processed in memory, but instead by the duration of memory processing.     

Is the interference between memory processing and timing specific to the use of verbal material?

Increasing load in a memory task performed simultaneously with a timing task shortens perceived time, an effect that has been observed previously with memory tasks using verbal material. The present experiments examine whether two similar memory tasks, one in which verbal material is used and another one in which nonverbal material is used, would produce similar interference effects on concurrent time reproduction. In Experiment 1, the number of nonverbal stimuli (pseudo-random dot patterns) was manipulated in a memory task performed while a temporal interval to be reproduced was encoded. Reproductions shortened proportionally to the duration of memory processing executed during time estimation. Verbal stimuli (consonants) were used in Experiment 2 in otherwise identical experimental conditions. Effects observed in Experiment 2 were comparable to those obtained in Experiment 1. Taken together, these results support the notion that interference from memory tasks on concurrent time estimation is not determined by the specific type of material processed in memory, but instead by the duration of memory processing.