Free explainer5 min read
Working memory
Understand the model, recognise it in a stem, separate the look-alikes, then apply it.
Start with the mental model
**Working memory is not a box for storage — it is a small mental workbench with four separate tool trays, and each tray can break independently of the others**. A person can hold a phone number perfectly (one tray intact) while being completely unable to picture a room's layout (a different tray broken) — and neither finding tells you anything about whether they can form a new long-term memory at all.
Visual explanationUse the diagram to rebuild the concept from memory.
01Core model
• Working memory, in Baddeley and Hitch's model, is an active system of four components. The CENTRAL EXECUTIVE allocates attention, switches strategy, and links to long-term memory. The PHONOLOGICAL LOOP holds verbal/acoustic information via a store plus rehearsal. The VISUOSPATIAL SKETCHPAD holds visual/spatial information, and the EPISODIC BUFFER binds information from the other components and long-term memory into one ordered representation.
• This replaced the older, unitary 'short-term memory' concept (Atkinson-Shiffrin), which treated brief storage as one passive box rather than an active system with specialised, separable parts.
• The phonological loop and visuospatial sketchpad dissociate on dual-task testing. Articulatory suppression (repeating an irrelevant sound aloud) disrupts verbal recall but not spatial recall. A competing spatial-tracking task disrupts spatial recall but not verbal recall instead — this is how the two-component split was originally demonstrated experimentally.
• The central executive is closely related to, but not the same as, executive function proper. The central executive allocates resources within working memory itself, while executive function is the broader system of planning, switching and inhibition that draws on working memory as one of its resources.
• Working memory is tested clinically by digit span (forward and backward) and by dual-task paradigms; a normal digit span reflects an intact phonological loop but says nothing about the visuospatial sketchpad, the episodic buffer, or any long-term memory system.
• In practice, this is why a patient can score normally on digit span testing yet still have severe, disabling amnesia — working memory and long-term episodic memory are entirely separate systems, and clinicians must test both rather than inferring one from the other.
02Recognise it in an SBA
• A vignette describing verbal recall disrupted by repeating an irrelevant sound aloud (articulatory suppression), with spatial recall unaffected, is describing the phonological loop specifically.
• A vignette describing normal digit span but an inability to hold a spatial layout or mentally rotate a shape is describing a selective visuospatial sketchpad deficit.
• A vignette describing difficulty switching attention between two working-memory subtasks, with broader real-world planning and social judgement intact, is describing the central executive, not executive function proper.
• A vignette describing binding of verbal, visual and background-knowledge content into one ordered mental representation held briefly is describing the episodic buffer, not episodic (long-term) memory.
• A vignette describing a normal digit span alongside a severe inability to form new long-term memories is demonstrating the working-memory/episodic-memory dissociationA disconnection between a person's thoughts, feelings, memories, or sense of identity, often as a response to stress.Read more → directly.
• A vignette describing preserved skills (riding a bicycle) despite dense amnesia is describing procedural memory, a long-term system unrelated to working-memory capacityA person's ability to understand, weigh up, and use relevant information to make a specific decision.Read more →.
03Separate the look-alikes
Working memory vs short-term memory — working memory is an active, four-component system that manipulates information; short-term memory (the older term) describes a single passive store without these separable parts. | |
Central executive vs executive function — the central executive allocates attention within working memory itself; executive function is the broader control system (planning, switching, inhibition) that uses working memory as one resource among several. | |
Phonological loop vs visuospatial sketchpad — verbal/acoustic content disrupted by articulatory suppression (phonological loop) vs visual/spatial content disrupted by a competing visual-tracking task (visuospatial sketchpad); they dissociate cleanly in both testing and lesion studies. | |
Episodic buffer vs episodic memory — the episodic buffer is a SHORT-TERM working-memory component that binds information briefly; episodic memory is a LONG-TERM, hippocampally-dependent system for recollecting past events — the shared word 'episodic' does not mean shared mechanism. | |
Working memory vs episodic memory — a normal digit span (working memory intact) does not rule out severe amnesia (episodic memory impaired); the two dissociate completely, most dramatically in classic amnesic syndromes. | |
Working memory vs semantic memory — semantic memory (long-term facts and concepts) and procedural memory (long-term skills) are further long-term systems, both independent of working memory capacityA person's ability to understand, weigh up, and use relevant information to make a specific decision.Read more →. | |
Working memory vs sustained attention — poor backward digit span can reflect a genuine central-executive/working-memory deficit OR a primary attention deficit; the two must be teased apart rather than assumed. |
How the exam thinksRead the question the way the examiner wrote it.
What people get wrong here
• ⚠️ Using 'working memory' and 'short-term memory' interchangeably as if they name the same model — they describe different theoretical frameworks (active multi-component system vs passive single store).
• Assuming a normal digit span means 'memory is fine' overall — it only speaks to the phonological loop, not to episodic, semantic or procedural long-term memory.
• Conflating the central executive with executive function proper because they share the word 'executive' — one is a working-memory subsystem, the other is the broader control system.
• Confusing the episodic buffer (a working-memory component) with episodic memory (a long-term system) purely because of the shared word 'episodic'.
• Attributing poor performance on a working-memory task automatically to a working-memory deficit without checking whether a primary attention problem is the real cause.
• Assuming a single working-memory deficit is global — a patient can have an isolated phonological-loop or visuospatial-sketchpad deficit with the other components, and all long-term systems, completely intact.