The multicomponent model of working memory fifty years on

Hitch, Graham J.; Allen, Richard J.; Baddeley, Alan D. · 2024 · Quarterly Journal of Experimental Psychology

DOI: 10.1177/17470218241290909

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This review article by Hitch, Allen, and Baddeley reflects on the evolution of the multicomponent model of working memory fifty years after its initial proposal. The authors address the original research question of whether short-term memory (STM) serves as a general-purpose working memory for cognition. Motivated by contradictory evidence—such as neuropsychological cases where STM deficits did not impair general cognition and experiments showing that disrupting STM did not affect long-term memory retrieval—the authors developed a dissociative methodology. By loading healthy adults’ STM with irrelevant information, they observed moderate interference in cognitive tasks like verbal reasoning and comprehension, suggesting that STM is only one part of a broader system. This led to the original model comprising a phonological loop for verbal storage and a central executive for attention-demanding control processes. Over the subsequent five decades, the model has been expanded and refined to increase its scope and applicability. The phonological loop was detailed through evidence of subvocal rehearsal and computational modeling of serial order, while the visuospatial sketchpad was added to account for visual and spatial storage, later distinguished into separable components. A major revision introduced the episodic buffer, a limited-capacity store for integrated, multimodal representations, to address the model’s initial neglect of long-term memory interactions and the need for binding disparate information. The authors describe their current view of the system as illustrated in Figure 1, where the episodic buffer acts as a central hub at the interface of perceptual input and internal control, while the central executive functions as a flexible resource for supervisory attention. The paper synthesizes findings regarding these components, noting that while the phonological loop’s exact decay mechanisms remain debated, its role in vocabulary acquisition is established. Research on the episodic buffer revealed that initial feature binding is largely automatic rather than executive-dependent, though maintenance involves attentional refreshing. The authors highlight that strategic prioritization can alter the schedule of this refreshing, boosting memory for high-value items. They also discuss how the model accommodates cross-modal integration and the influence of long-term memory on working memory performance. Comparisons with alternative models suggest broad agreement on the phenomena to be explained, with more similarities than differences. The significance of this review lies in its assertion that the longevity of the multicomponent model stems from its simplicity, robustness, and utility as a framework for further theoretical development. The authors conclude that while differences between models attract interest, the most critical issues for future research concern the nature of executive control. They emphasize that the model provides a high-level map of the working memory system, facilitating the investigation of processes and mechanisms even as lower-level details, such as exact neural boundaries, remain to be fully understood. The framework continues to offer practical value in understanding cognition and clinical disorders.

Key finding

The multicomponent working memory model's core dissociations — between attentional/executive resources and domain-specific temporary storage, and between verbal (phonological loop) and visuospatial (sketchpad) buffers — have remained robust across fifty years of dual-task and neuropsychological evidence. Differences between competing working memory models reflect emphasis more than fundamental disagreement; the principal open question is the nature of executive control.

Methodology

review

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via open_access_repository on 2026-05-07 (5 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• working memory
• capacity resource theory
• cognitive capacity variation