Models of verbal working memory capacity: what does it take to make them work?

Theories of working memory (WM) capacity limits will be more useful when we know what aspects of performance are governed by the limits and what aspects are governed by other memory mechanisms. Whereas considerable progress has been made on models of WM capacity limits for visual arrays of separate objects, less progress has been made in understanding verbal materials, especially when words are mentally combined to form multiword units or chunks. Toward a more comprehensive theory of capacity limits, we examined models of forced-choice recognition of words within printed lists, using materials designed to produce multiword chunks in memory (e.g., leather brief case). Several simple models were tested against data from a variety of list lengths and potential chunk sizes, with test conditions that only imperfectly elicited the interword associations. According to the most successful model, participants retained about 3 chunks on average in a capacity-limited region of WM, with some chunks being only subsets of the presented associative information (e.g., leather brief case retained with leather as one chunk and brief case as another). The addition to the model of an activated long-term memory component unlimited in capacity was needed. A fixed-capacity limit appears critical to account for immediate verbal recognition and other forms of WM. We advance a model-based approach that allows capacity to be assessed despite other important processing contributions. Starting with a psychological-process model of WM capacity developed to understand visual arrays, we arrive at a more unified and complete model.     

How does running memory span work?

In running memory span, a list ends unpredictably, and the last few items are to be recalled. This task is of increasing importance in recent research. We argue that there are two very different strategies for performing running span tasks: a low-effort strategy in which items are passively held until the list ends, when retrieval into a capacity-limited store takes place; and a higher-effort strategy in which working memory is continually updated using rehearsal processes during the list presentation. In two experiments, we examine the roles of these two strategies and the consequences of two types of interference.     

How does running memory span work?

In running memory span, a list ends unpredictably, and the last few items are to be recalled. This task is of increasing importance in recent research. We argue that there are two very different strategies for performing running span tasks: a low-effort strategy in which items are passively held until the list ends, when retrieval into a capacity-limited store takes place; and a higher-effort strategy in which working memory is continually updated using rehearsal processes during the list presentation. In two experiments, we examine the roles of these two strategies and the consequences of two types of interference.     

Does working memory training work? The promise and challenges of enhancing cognition by training working memory

A growing body of literature shows that one's working memory (WM) capacity can be expanded through targeted training. Given the established relationship between WM and higher cognition, these successful training studies have led to speculation that WM training may yield broad cognitive benefits. This review considers the current state of the emerging WM training literature, and details both its successes and limitations. We identify two distinct approaches to WM training, strategy training and core training, and highlight both the theoretical and practical motivations that guide each approach. Training-related increases in WM capacity have been successfully demonstrated across a wide range of subject populations, but different training techniques seem to produce differential impacts upon the broader landscape of cognitive abilities. In particular, core WM training studies seem to produce more far-reaching transfer effects, likely because they target domain-general mechanisms of WM. The results of individual studies encourage optimism regarding the value of WM training as a tool for general cognitive enhancement. However, we discuss several limitations that should be addressed before the field endorses the value of this approach.     

Is working memory still working?

The current state of A. D. Baddeley and G. J. Hitch's (1974) multicomponent working memory model is reviewed. The phonological and visuospatial subsystems have been extensively investigated, leading both to challenges over interpretation of individual phenomena and to more detailed attempts to model the processes underlying the subsystems. Analysis of the controlling central executive has proved more challenging, leading to a proposed clarification in which the executive is assumed to be a limited capacity attentional system, aided by a newly postulated fourth system, the episodic buffer. Current interest focuses most strongly on the link between working memory and long-term memory and on the processes allowing the integration of information from the component subsystems. The model has proved valuable in accounting for data from a wide range of participant groups under a rich array of task conditions. Working memory does still appear to be working.     

Does visual working memory work as a few fixed slots?

The present study investigated whether visual working memory (VWM) functions as a few (about 3?～ 4) fixed slots by examining how the distribution of VWM is adjusted. Adopting a change-detection paradigm, we required subjects to memorize four items, one of which was prioritized. If VWM functions as 3?～?4 slots, allocating multiple slots to the prioritized item would leave no slot for some other items; consequently no information would be stored for them, leading to a substantial decrease in change-detection performance no matter whether small or large changes occurred. The result showed that small changes on the unfavoured items were detected less accurately, indicating that more VWM was allocated to the favoured item. Yet meanwhile large changes that occurred on those unfavoured items could still be detected as well as those on the favoured one, indicating that each of them was still stored to some extent rather than completely discarded. The results suggested that VWM may not work as 3?～?4 fixed slots. Possible mechanisms were discussed based on the present results, including a modified slot model with more available slots, a continuous resource model, and a hierarchical model that assumes storage of ensemble information in addition to the information of individual items.     

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.     

Developmental differences in working memory: Where do they come from?

Several models assume that working memory development depends on age-related increases in efficiency and speed of processing. However, age-related increases in the efficiency of the mechanisms that counteract forgetting and restore memory traces may also be important. This hypothesis was tested in three experiments by manipulating both the processing duration within a working memory task and the time available to restore memory traces. Third- and sixth-grade children performed a complex span task in which they maintained series of letters while adding numbers to series of digits. When we equated processing and restoration times between ages, the developmental difference in working memory span was reduced but remained significant. However, this residual difference was eliminated when the time available to reactivate memory traces was tailored to the processing speed of each age group. This indicates that children employ active mechanisms for maintenance and restoration of memory traces that develop with age.     

Variation in working memory capacity and intrusions: Differences in generation or editing?

The current study explored the reason for the relation between individual differences in working memory capacity (WMC) and intrusions in free recall. High and low WMC individuals were tested in standard delayed free recall and externalised free recall in which participants recalled everything that came to mind. Additionally, in externalised free recall participants were instructed to press a key for each item that they knew was an intrusion. In delayed free recall, low WMC individuals recalled fewer correct items and more previous list and extralist intrusions than high WMC individuals. In externalised free recall, differences only arose in previous list intrusions. Furthermore, in externalised free recall it was found that low WMC were less likely to identify both types of intrusions than high WMC individuals. It is argued that the reason low WMC individuals recall more intrusions than high WMC in free recall is due to differences in both generation and editing abilities.     

Increased distractibility in schizotypy: Independent of individual differences in working memory capacity?

Individuals with schizophrenia typically show increased levels of distractibility. This has been attributed to impaired working memory capacity (WMC), since lower WMC is typically associated with higher distractibility, and schizophrenia is typically associated with impoverished WMC. Here, participants performed verbal and spatial serial recall tasks that were accompanied by to-be-ignored speech tokens. For the few trials wherein one speech token was replaced with a different token, impairment was produced to task scores (a deviation effect). Participants subsequently completed a schizotypy questionnaire and a WMC measure. Higher schizotypy scores were associated with lower WMC (as measured with operation span, OSPAN), but WMC and schizotypy scores explained unique variance in relation to the mean magnitude of the deviation effect. These results suggest that schizotypy is associated with heightened domain-general distractibility, but that this is independent of its relationship with WMC.     

Do objects in working memory compete with objects in perception?

It is generally assumed that “perceptual object” is the basic unit for processing visual information and that only a small number of objects can be either perceptually selected or encoded in working memory (WM) at one time. This raises the question whether the same resource is used when objects are selected and tracked as when they are held in WM. In two experiments, we measured dual-task interference between a memory task and a Multiple Object Tracking task. The WM tasks involve explicit, implicit, or no spatial processing. Our results suggest there is no resource competition between working memory and perceptual selection except when the WM task requires encoding spatial properties.     

Is working memory training effective?

Working memory (WM) is a cognitive system that strongly relates to a person's ability to reason with novel information and direct attention to goal-relevant information. Due to the central role that WM plays in general cognition, it has become the focus of a rapidly growing training literature that seeks to affect broad cognitive change through prolonged training on WM tasks. Recent work has suggested that the effects of WM training extend to general fluid intelligence, attentional control, and reductions in symptoms of ADHD. We present a theoretically motivated perspective of WM and subsequently review the WM training literature in light of several concerns. These include (a) the tendency for researchers to define change to abilities using single tasks, (b) inconsistent use of valid WM tasks, (c) no-contact control groups, and (d) subjective measurement of change. The literature review highlights several findings that warrant further research but ultimately concludes that there is a need to directly demonstrate that WM capacity increases in response to training. Specifically, we argue that transfer of training to WM must be demonstrated using a wider variety of tasks, thus eliminating the possibility that results can be explained by task specific learning. Additionally, we express concern that many of the most promising results (e.g., increased intelligence) cannot be readily attributed to changes in WM capacity. Thus, a critical goal for future research is to uncover the mechanisms that lead to transfer of training.     

Multiple-target tracking: A role for working memory?

In order to identify the cognitive processes associated with target tracking, a dual-task experiment was carried out in which participants undertook a dynamic multiple-object tracking task first alone and then again, concurrently with one of several secondary tasks, in order to investigate the cognitive processes involved. The research suggests that after designated targets within the visual field have attracted preattentive indexes that point to their locations in space, conscious processes, vulnerable to secondary visual and spatial task interference, form deliberate strategies beneficial to the tracking task, before tracking commences. Target tracking itself is realized by central executive processes, which are sensitive to any other cognitive demands. The findings are discussed in the context of integrating dynamic spatial cognition within a working memory framework.     

Nonverbal working memory of humans and monkeys: Rehearsal in the sketchpad?

Investigations of working memory tend to focus on the retention of verbal information. The present experiments were designed to characterize the active maintenance rehearsal process used in the retention of visuospatial information. Rhesus monkeys (Macaca mulatta;N=6) were tested as well as humans (totalN=90) because these nonhuman primates have excellent visual working memory but, unlike humans, cannot verbally recode the stimuli to employ verbal rehearsal mechanisms. A series of experiments was conducted using a distractor-task paradigm, a directed forgetting procedure, and a dual-task paradigm. No evidence was found for an active maintenance process for either species. Rather, it appears that information is maintained in the visuospatial sketchpad without active rehearsal.     

Evidence for spontaneous serial refreshing in verbal working memory?

Working memory (WM) keeps information temporarily accessible for ongoing cognition. One proposed mechanism to keep information active in WM is refreshing. This mechanism is assumed to operate by bringing memory items into the focus of attention, thereby serially refreshing the content of WM. We report two experiments in which we examine evidence for the spontaneous occurrence of serial refreshing in verbal WM. Participants had to remember series of red letters, while black probe letters were presented between these memory items, with each probe to be judged present in or absent from the list presented so far, as quickly as possible (i.e., the probe–span task). Response times to the probes were used to infer the status of the representations in WM and, in particular, to examine whether the content of the focus of attention changed over time, as would be expected if serial refreshing occurs spontaneously during inter-item pauses. In sharp contrast with this hypothesis, our results indicate that the last-presented memory item remained in the focus of attention during the inter-item pauses of the probe–span task. We discuss how these findings help to define the boundary conditions of spontaneous refreshing of verbal material in WM, and discuss implications for verbal WM maintenance and forgetting.     

How does the severity of a learning disability affect working memory performance?

Working memory performance was examined in children aged 11-12 years who had borderline, mild, and moderate learning disabilities. Comparisons with children of average abilities were used to determine whether those with more severe learning disabilities had greater impairments in working memory. Seven measures of working memory span were used to assess temporary phonological short-term storage (digit span, word span), temporary visuo-spatial short-term storage (pattern span, spatial span), and temporary short-term storage with additional processing, or central executive, demands (listening span, odd one out span, reverse digit span). Children with mild and moderate learning disabilities were impaired on all measures of working memory compared to children of average abilities. Children with borderline learning disabilities were just as good as children with average abilities on visuo-spatial and complex span tasks, but showed an impairment on phonological span tasks. Children with moderate learning disabilities were indistinguishable from children with mild learning disabilities on simple span tasks, but were significantly poorer than the mild group on the more demanding complex span tasks. For the group as a whole, working memory was strongly related to mental age.