What limits working memory capacity?

Abstract

We review the evidence for the three principal theoretical contenders that vie to explain why and how working memory capacity is limited. We examine the possibility that capacity limitations arise from temporal decay; we examine whether they might reflect a limitation in cognitive resources; and we ask whether capacity might be limited because of mutual interference of representations in working memory. We evaluate each hypothesis against a common set of findings reflecting the capacity limit: The set-size effect and its modulation by domain-specificity and heterogeneity of the memory set; the effects of unfilled retention intervals and of distractor processing in the retention interval; and the pattern of correlates of working-memory tests. We conclude that---at least for verbal memoranda---a decay explanation is untenable. A resource-based view remains tenable but has difficulty accommodating several findings. The interference approach has its own set of difficulties but accounts best for the set of findings, and therefore appears to present the most promising approach for future development.Keywords: working memory, decay, resources, interference, individual differences WM Capacity 3 What Limits Working Memory Capacity?Working memory (WM) is the system that holds mental representations available for processing. Its limited capacity is a limiting factor for the complexity of our thoughts (Halford, Cowan, & Andrews, 2007;Oberauer, 2009). Measures of WM capacity have been identified as major determinants of cognitive development in childhood (Bayliss, Jarrold, Gunn, & Baddeley, 2003) and in old age (Park et al., 2002;Salthouse, 1994), as well as of individual differences in intellectual abilities (Conway, Kane, & Engle, 2003;Jarrold & Towse, 2006). Understanding why WM capacity is limited is therefore an essential step toward understanding why human cognitive abilities are limited, why individuals differ in these abilities, and how abilities develop over the lifespan.In this article we use the term WM capacity in a descriptive sense, referring to the fact that people can hold only a limited amount of mental content available for processing. The capacity limit is usually operationalized as a limit on how much new information people can remember over short periods of time (in the order of seconds), but there are reasons to believe (discussed below) that the capacity limit also applies to people's ability to make information in the current environment simultaneously available for processing.Hypotheses about what limits WM capacity can be organized into three groups: (1) Some theories assume that representations in WM decay over time, unless decay is prevented by some form of restoration process such as rehearsal. According to this view, WM has limited capacity because only a limited amount of information can be rehearsed before it fades away into an irrecoverable state (Baddeley, Thomson, & Buchanan, 1975;Schweickert & Boruff, 1986). (2) Alternatively, WM capacity has been characterized as a limited resourc...

Shared processing in multiple object tracking and visual working memory in the absence of response order and task order confounds

“…Our data show significant individual differences in performance, which has been a hallmark of the literature to date and has proven useful in the theoretical analysis of capacity limitations in working memory [25], though it is often obscured through the presentation of group data. When we examine the individual-subject results, we progressively reduce the differences in performance between subjects through our experimental progression by calibrating to individual baseline performance (1a-1b) and removing stimulus-based confounds in the experimental paradigm (1b-2).…”

“…That is not to say that these alternative theories do not have a role in our overall understanding of the processes of memory, nor that they do not influence one another in their respective theoretical progression, as they clearly do [2]. It is also the case that more general theories of attention and integration of information must be involved to provide a full account of the underlying processes [35–38], and understanding the complex nature of working memory in all its forms remains an ongoing challenge [25]. In light of this, and for the purpose of the current discussion of what is primarily an empirical paper, we will focus on the M-WM [1, 2, 26, 27] as has the majority of the immediately relevant literature [4–7].…”

Frequency Effects on Memory: A Resource-Limited Theory

“…Numerous studies have shown that it is not the absolute amount of processed information that limits WM capacity, but it is rather the number of chunks into which this information can be compressed and organized (Miller, 1956;Simon, 1974; for recent reviews, see Cowan, 2001;or Gobet et al, 2001). Researchers may disagree about the specific number of units that constitute the limit of WM capacity (Cowan, 2001;Gobet & Clarkson, 2004;Luck & Vogel, 1997;Miller, 1956) or about the mechanism responsible (Cowan, 2001;Gobet et al, 2001;Oberauer et al, 2016); however, most theories of WM treat chunks as all or nothing. Current verbal learning theories do not see chunks as varying in strength, and, as a consequence, they do not consider that variation in chunk strength would affect WM capacity 11 .…”

“…Multinomial modeling has revealed that WM capacity correlates with encoding success, rather than with retrieval probability (Marevich et al, 2017). This correlation is one of the key puzzles that a theory of WM must explain, and Oberauer et al (2016) noted that no current resource-based account of WM can do so. The theory presented in this paper provides a natural account of for this observation -the creation and binding of novel episode nodes depletes a limited resource, and the strength or the probability of forming these nodes depends on the quantity of WM resources available.…”

“…As a result, few memory models posit a role for prior item strength in episodic learning, knowledge formation, or WM capacity (for a review of WM models, see Oberauer, Farrell, Jarrold, & Lewandowsky, 2016).…”