Central Executive vs. Episodic Buffer

The architecture of human cognition relies fundamentally on our capacity to maintain, manipulate, and synthesize information across transient temporal windows. In cognitive psychology and neuroscience, this apparatus is conceptualized as working memory—a limited-capacity system that underpins reasoning, comprehension, learning, and goal-directed behavior. For decades, the dominant theoretical framework guiding working memory research has been the multicomponent model originally proposed by Alan Baddeley and Graham Hitch. Within this paradigm, cognitive functions are not subsumed under a single, monolithic short-term storage bin; instead, they are fractionated into specialized subsystems operating under coordinated executive oversight.

Among the components of this cognitive architecture, two structures occupy central theoretical importance: the Central Executive and the Episodic Buffer. The Central Executive serves as the supervisory attentional controller, dictating the allocation of cognitive resources, scheduling task-switching routines, and suppressing prepotent distractions without maintaining information within an internal representational medium. Conversely, the Episodic Buffer—introduced decades later to resolve glaring empirical anomalies—functions as a bounded, multimodal storage reservoir capable of synthesizing disparate sensory features, linguistic constructs, and long-term memory traces into coherent, chronologically organized episodic representations.

Despite their distinct theoretical roles, students, researchers, and clinicians frequently conflate the Central Executive and the Episodic Buffer. This confusion arises because both operate at the apex of conscious cognitive processing, both interface intimately with long-term memory systems, and both are recruited during complex, high-demand operations such as narrative comprehension, mental arithmetic, and creative problem-solving. This analysis examines the functional dissociations, neuroanatomical substrates, operational dynamics, and theoretical evolutions that differentiate the Central Executive from the Episodic Buffer, clarifying how human working memory balances supervisory control with multimodal integration.

1. Introduction to Working Memory Architecture

1.1 Evolution of Baddeley’s Multicomponent Model

The contemporary understanding of working memory emerged as a direct theoretical departure from the unitary short-term memory models that dominated the mid-twentieth century. Most notably, the modal model advanced by Richard Atkinson and Richard Shiffrin in 1968 conceptualized memory as a linear progression of structures: sensory registers transferred information into a single short-term store (STS), which simultaneously functioned as a temporary working memory and the mandatory gateway through which representations entered long-term storage (LTS). While parsimonious, the Atkinson-Shiffrin model soon encountered insurmountable empirical challenges. Neuropsychological investigations in the late 1960s and early 1970s revealed patients who displayed profound deficits in short-term digit span tasks yet retained unimpaired long-term learning capabilities and normal higher-order cognitive processing. Such clinical profiles contradicted the hypothesis that a unitary short-term store served as the obligatory bottleneck for long-term encoding and complex reasoning.

In response to these empirical discrepancies, Alan Baddeley and Graham Hitch (1974) formulated the tripartite model of working memory. Rather than viewing temporary memory as a passive, single-chamber reservoir, Baddeley and Hitch envisioned a modular, dynamic system comprising an attentional supervisor and specialized maintenance subsystems. At its inception, the tripartite model was defined by three distinct components:

• The Central Executive: A domain-general processing engine responsible for regulating information flow, allocating attentional capacity, and directing strategic mental operations.
• The Phonological Loop: A modality-specific slave system dedicated to the maintenance of acoustic and verbal material through a passive phonological store and an active subvocal articulatory rehearsal mechanism.
• The Visuospatial Sketchpad: A parallel modality-specific slave system tasked with retaining and manipulating visual imagery, spatial coordinates, and movement sequences.

While the tripartite model successfully accounted for a wide range of dual-task interference patterns, its architectural limitations became increasingly problematic over the subsequent quarter-century. By isolating verbal and visuospatial maintenance into non-communicating slave systems governed by a non-storing central executive, the 1974 model could not account for cross-modal integration. Real-world cognition rarely operates in strictly segregated sensory channels; individuals routinely bind the visual appearance of an object to its spoken name, or remember the spatial trajectory of an auditory signal.

Furthermore, the tripartite architecture failed to accommodate the substantial influence of long-term memory on immediate recall spans. For example, individuals routinely demonstrate a memory span of fifteen to twenty words when those words are presented as a syntactically coherent sentence, despite having an immediate span of merely five to six unrelated words. The phonological loop lacked the informational capacity to hold such extensive semantic sequences, yet the central executive explicitly lacked storage capacity. Recognizing that the original model possessed no theoretical mechanism to integrate cross-modal tokens or to bind immediate perceptual data with long-term semantic structures, Baddeley (2000) introduced a structural revision: the Episodic Buffer. This fourth component transformed working memory from a set of modular loops into an integrated cognitive workspace capable of multimodal representation.

1.2 Core Definitions and Causes of Confusion

To establish conceptual clarity, it is necessary to construct rigorous, non-overlapping operational definitions for both the Central Executive and the Episodic Buffer, while deconstructing the historical and structural reasons for their persistent conflation in cognitive psychology literature.

The Central Executive (CE) is defined as a domain-general supervisory attentional system. Its operational parameters are exclusively processor-oriented. It possesses zero internal mnemonic storage capacity in all formulations following Baddeley’s adoption of the Supervisory Attentional System framework in 1986. The CE regulates mental operations by focusing attentional resources, dividing attention between simultaneous streams of sensory input, shifting cognitive mindsets across dynamic task environments, and deliberately inhibiting automated, prepotent behavioral routines. It does not contain sensory traces, verbal codes, visual images, or semantic propositions; rather, it directs the neurological and cognitive machinery that manipulates those representations.

The Episodic Buffer (EB) is defined as a limited-capacity, multimodal storage module. Its operational parameters are primarily representational. It possesses a capacity ceiling empirically circumscribed at approximately four multidimensional chunks of integrated information. The EB relies upon a common multidimensional code capable of translating and binding the heterogeneous outputs of the phonological loop, the visuospatial sketchpad, and long-term memory into cohesive, unitary episodes. While passive or semi-passive in its baseline storage characteristics, it preserves the temporal sequence of events, providing a coherent chronological record of immediate conscious experience.

Conceptual confusion between these two modules typically stems from three primary sources:

1. Intimate Coupling in Conscious Awareness: Both the Central Executive and the Episodic Buffer operate in close connection with conscious phenomenological experience. The items occupying the focus of attention directed by the Central Executive correspond closely to the integrated multidimensional tokens maintained within the Episodic Buffer. Because conscious reflection on a memory trace often occurs concurrently with the executive manipulation of that trace, observers readily mistake the storage vessel (the buffer) for the attentional agent inspecting it (the executive).
2. Historical Conflation of Excess Capacity: In the decades preceding the year 2000, experimental psychologists frequently attributed unexplained performance variance—such as immediate prose recall, semantic chunking, and cross-modal binding—to the “residual capacity” of the Central Executive. Because early conceptualizations vaguely permitted the executive to double as a general work space, many empirical findings that actually demonstrated episodic buffering were cataloged under the umbrella of executive control.
3. The Dual-Task Degradation Fallacy: Experimental designs attempting to isolate working memory components often employ secondary executive loading tasks (such as random number generation or continuous backward counting). When these secondary tasks disrupt cross-modal binding or immediate narrative memory, researchers historically concluded that the storage itself was an executive function. Modern investigations have revealed that while the coordination and active retrieval of bound representations may recruit executive bandwidth, the representational storage occurs within the structurally distinct episodic buffer.

The functional boundary is straightforward: the Central Executive provides the dynamic processing, selection, and inhibition necessary for goal-directed control, whereas the Episodic Buffer provides the temporary multimodal representational workspace necessary to hold the products of that processing.

2. Overview of the Central Executive

2.1 Theoretical Foundations and Norman-Shallice Architecture

The conceptual framework of the Central Executive was significantly refined when Alan Baddeley (1986) formally integrated the attentional control model developed by Donald Norman and Tim Shallice. The Norman-Shallice model posits a two-tiered architecture governing human action selection and cognitive control: Contention Scheduling and the Supervisory Attentional System (SAS).

Contention Scheduling is an automatic, decentralized mechanism that resolves conflicts among routine, highly practiced action schemas. In familiar environments, perceptual inputs automatically activate associated behavioral schemas. When multiple routine actions compete for execution—such as steering an automobile while simultaneously operating the turn indicator—contention scheduling uses lateral inhibition to allow the dominant schema to execute smoothly without conscious, deliberative intervention. This lower-tier process operates with high computational efficiency and negligible consumption of conscious attentional resources.

However, contention scheduling fails when an organism encounters novel circumstances, high-stakes decision points, complex problem-solving requirements, or situations where a strongly habitual response must be suppressed in favor of an unfamiliar action. To address these demands, Norman and Shallice posited the Supervisory Attentional System (SAS), a top-down, non-routine control mechanism. The SAS can inject top-down activation or inhibition directly into contention scheduling networks, overriding default automated routines and deliberately biasing schema selection toward internally maintained behavioral goals.

Baddeley mapped the Central Executive directly onto the theoretical architecture of the SAS. Under this operationalization, the Central Executive is not a repository for perceptual information; it is the attentional coordinator that instantiates top-down bias. It monitors environmental feedback, evaluates the success of active behavioral plans, resolves interference among competing inputs, and dynamically reallocates finite processing resources across the peripheral cognitive subsystems. When an individual must deliberately navigate an unfamiliar detour while driving through a severe storm, the Central Executive intervenes, suppressing conversation (inhibiting the phonological loop) and concentrating all perceptual processing capacity onto immediate spatial navigation (the visuospatial sketchpad).

2.2 Core Cognitive Operations

Rather than treating the Central Executive as a monolithic entity, contemporary cognitive science—influenced by the seminal fractionation work of Akira Miyake and colleagues (2000) as well as Baddeley’s own experimental programs—deconstructs executive functioning into four interrelated core cognitive operations: selective focused attention, task-switching and cognitive flexibility, divided attention, and the inhibition of prepotent responses.

Selective and Focused Attention: This operation represents the ability to direct cognitive resources toward specific task-relevant sensory streams or internal mental states while systematically filtering out irrelevant environmental distractors. Selective attention acts as an informational filter, ensuring that sensory inputs do not overwhelm the limited capacities of working memory. Neurobiologically, this process involves top-down amplification of task-relevant sensory signals within early perceptual cortices, paired with active suppression of background noise.

Task-Switching and Cognitive Flexibility: Real-world cognition requires constant adaptation to fluctuating environmental demands. Shifting involves disengaging attentional focus from an active mental schema, rule set, or behavioral objective and immediately establishing a new, alternative schema. This process carries a quantifiable behavioral penalty termed the “switch cost”—manifested as increased reaction time and elevated error rates immediately following a transition. The Central Executive oversees the reconstruction of internal task goals, the reconfiguration of perceptual processing filters, and the retrieval of appropriate task-specific rules from long-term memory.

Divided Attention: While selective attention requires the isolation of a single stream, divided attention involves the strategic coordination and concurrent distribution of processing capacity across two or more non-interchangeable tasks. Through dual-task experimental paradigms—such as requiring a participant to track a moving visual target while simultaneously performing continuous auditory verbal generation—researchers have demonstrated that the Central Executive acts as a temporal scheduler. Rather than executing true parallel processing for complex, capacity-demanding tasks, the executive coordinates micro-timesharing between competing operational streams, balancing processing bottlenecks to minimize operational degradation.

Inhibition of Prepotent Responses: This operation involves the intentional suppression of automated, dominant, or reflexive behavioral outputs that are inappropriate for the current goal state. Prototypically indexed by the Stroop task, the antisaccade task, and Go/No-Go paradigms, inhibitory control requires the Central Executive to withhold a habitual motor action or linguistic response (such as reading an ink color rather than the word itself). Inhibitory control also protects working memory buffers from proactive interference, actively purging outdated representations that could otherwise compromise the encoding of novel information.

2.3 Neuroanatomical Correlates

The cognitive operations comprising the Central Executive do not reside within a single, anatomically isolated brain center; instead, they emerge from the synchronized dynamics of a distributed, large-scale frontoparietal system commonly designated as the Central Executive Network (CEN).

The primary cortical hub of the Central Executive is the dorsolateral prefrontal cortex (dlPFC), encompassing Brodmann Areas (BA) 9 and 46. The dlPFC provides the neurobiological infrastructure for active goal maintenance, rule representation, working memory monitoring, and top-down attentional biasing. Neurons within the dlPFC exhibit sustained, elevated firing rates across delay intervals during working memory tasks, maintaining representations of behavioral rules and task schemas in an active state even in the presence of strong sensory distractors. The dlPFC acts as the operational conductor, sending top-down regulatory projections to posterior sensory cortices, motor planning regions, and subcortical nuclei.

A second neuroanatomical component is the anterior cingulate cortex (ACC), specifically its dorsal division (BA 24/32). The ACC does not execute attentional control directly; rather, it functions as an evaluative performance monitoring engine. It detects cognitive conflict, operational errors, and unfavorable outcomes. When the ACC identifies simultaneous activation of mutually incompatible response tendencies (such as during the incongruent condition of a Stroop task), it signals the dlPFC to recruit higher levels of top-down inhibitory control, dynamically resolving the processing conflict.

The frontoparietal architecture of the CEN is completed by the posterior parietal cortex, with particular emphasis on the intraparietal sulcus (IPS) and the superior parietal lobule (SPL) (BA 7/40). The parietal components of this network coordinate the spatial allocation of attention, maintain task-relevant attentional priority maps, and facilitate the structural translation of abstract executive goals into concrete perceptual selections.

The clinical importance of this frontoparietal circuitry is demonstrated in dysexecutive syndrome, which typically emerges following focal lesions to the prefrontal cortex resulting from traumatic brain injury, ischemic stroke, or frontotemporal lobar degeneration. Patients exhibiting dysexecutive syndrome frequently demonstrate preserved perceptual abilities, normal speech production, and intact performance on conventional single-modality short-term memory tests (such as forward digit span). However, they manifest catastrophic failures in cognitive flexibility, abstract planning, problem-solving, and impulse regulation. In experimental settings, these patients display perseveration—continually executing an outdated rule set on the Wisconsin Card Sorting Test despite clear feedback that their choices are incorrect—and severe distractibility, underscoring the absolute breakdown of the Supervisory Attentional System.

2.4 Theoretical Limitations and the Homunculus Problem

Despite its utility in organizing cognitive research, the Central Executive has historically drawn sharp criticism from theoretical psychologists and philosophers of mind regarding the “homunculus problem.” In psychological modeling, a homunculus—literally a “little person inside the head”—represents an explanatory fallacy wherein an unmodeled, internally autonomous agent is invoked to explain complex behaviors, merely pushing the explanatory requirement back one level without addressing how the underlying mechanism functions.

In early descriptions of the multicomponent model, the Central Executive was vulnerable to this critique. Whenever human participants made strategic decisions, resolved cognitive dilemmas, prioritized tasks, or retrieved long-term information in ways that peripheral slave systems could not accommodate, the behavior was attributed to the Central Executive. Critics pointed out that labeling an internal process as an executive action does not illuminate the computational mechanisms producing that action; it merely invents an unobservable conductor executing those decisions.

Alan Baddeley acknowledged this conceptual vulnerability, famously stating that the homunculus was a necessary theoretical placeholder—a conceptual convenience marking the boundaries of what remained scientifically unexplained within working memory. To dismantle the homunculus, Baddeley and contemporary cognitive researchers engaged in systematic fractionation: deconstructing the executive into distinct, mathematically and empirically falsifiable sub-processes. By decomposing supervisory control into specific mechanisms—such as the set-shifting metrics formalised by Miyake, conflict-monitoring computational models of the ACC, and top-down gating mechanisms—researchers have progressively replaced homuncular agency with neurobiologically plausible networks.

Nevertheless, a lingering structural limitation remained within the classic formulation of the Central Executive: its absolute lack of storage capacity. By defining the CE strictly as a processing controller without representational real estate, cognitive theory was left without a mechanism to explain where the products of executive synthesis were maintained during complex reasoning. If the Central Executive could not hold information, and the peripheral slave systems could only retain unimodal auditory or visuospatial tokens, the multicomponent model lacked an integrated workspace where active processing and multimodal storage could coalesce. This theoretical gap ultimately forced the conceptualization and structural addition of the Episodic Buffer.

3. Overview of the Episodic Buffer

3.1 Historical Context and Justification for Inclusion

The structural integration of the Episodic Buffer into Baddeley’s multicomponent model in the year 2000 was necessitated by an accumulation of empirical findings that the tripartite model could neither predict nor explain. For nearly three decades, working memory research had operated under the premise that storage was entirely segregated into modality-specific buffers (the phonological loop for acoustic-verbal sequences and the visuospatial sketchpad for spatial-visual features), while the central executive performed operations without storing data. However, several clinical and experimental findings rendered this architecture untenable.

The most prominent catalyst was the phenomenon of immediate sentence and prose recall. Under standard experimental conditions, an individual’s immediate verbatim recall span for completely unrelated words hovers between five and six items, constrained by the temporal decay and capacity boundaries of the phonological loop. However, when those exact same lexical items are arranged into a syntactically correct, semantically meaningful sentence, an individual’s immediate recall span surges to fifteen, sixteen, or even twenty words. If the phonological loop has a fixed capacity governed by articulatory duration, and the central executive possesses zero storage capacity, where was this vast quantity of linguistic and semantic information maintained during immediate recall?

A second empirical challenge emerged from neuropsychological case studies of densely amnesic patients. Clinical studies, including investigations by Baddeley and Barbara Wilson (2002), focused on patients suffering from bilateral medial temporal lobe damage. These individuals had severe anterograde amnesia and were entirely incapable of encoding new episodic traces into long-term memory. Yet, when presented with complex, multi-sentence narrative passages and tested for immediate recall within seconds, these patients demonstrated preserved recall of rich, multifaceted stories. They retained not merely individual phonological items, but the intricate conceptual associations, semantic linkages, and temporal trajectories of the narrative. Because their long-term episodic memory system was destroyed, this information could not be stored in long-term memory; and because the narrative far exceeded the capacity of the phonological loop, it could not reside within unimodal slave systems. The preserved performance indicated the existence of a distinct, short-term, limited-capacity, multimodal episodic storage space.

A third anomaly involved the cognitive demand of cross-modal feature binding. In the real world, human experience is intrinsically multimodal. In everyday tasks, visual inputs (such as reading text) immediately trigger auditory representations (subvocal speech), which in turn activate complex semantic concepts retrieved from long-term memory. The original tripartite model had no direct informational bridge between the phonological loop and the visuospatial sketchpad; information could only theoretically communicate through the non-storing central executive. To address these theoretical gaps, Baddeley formally introduced the Episodic Buffer as a third slave system possessing passive storage properties and a common multimodal representational code.

3.2 Structural Mechanisms and Information Binding

The Episodic Buffer serves as an integrated representational interface within working memory. Its structural mechanisms are governed by three primary characteristics: capacity limits, a common multimodal code, and chronological sequencing.

Capacity Limits: Unlike the phonological loop, whose capacity is constrained by temporal duration (approximately two seconds of subvocal speech), the Episodic Buffer is capacity-constrained by the number of integrated informational units it can support. Empirical evidence aligns the buffer’s capacity with the general working memory capacity limit identified by Nelson Cowan (2001): approximately four multidimensional chunks. A chunk is an integrated informational packet; it may consist of a single visual object, an interconnected semantic phrase, or a multifaceted mental scene containing visual, spatial, and acoustic features.

The Multidimensional Representational Code: The defining architectural property of the Episodic Buffer is its capacity to operate using an integrated, multidimensional representational format. The peripheral slave systems communicate in sensory-specific codes: the phonological loop uses acoustic, phonemic, and temporal articulatory signals, while the sketchpad uses categorical spatial coordinates and retinotopic visual features. The Episodic Buffer acts as a common translation platform. It projects these heterogeneous streams into a shared abstract medium, allowing visual tokens (e.g., the visual appearance of a dog) to bind seamlessly with auditory signals (e.g., the sound of a bark), spatial coordinates (e.g., running in the yard), and semantic concepts (e.g., pet ownership retrieved from long-term memory).

Temporal Coherence and Chronological Sequencing: Beyond cross-modal binding, the buffer is explicitly designated as “episodic” because it preserves temporal structure. Human experience does not unfold as disconnected sensory impressions; it is perceived as an orderly sequence of events over time. The Episodic Buffer organizes incoming bound features chronologically, maintaining an immediate running narrative of perceptual reality. This sequencing function enables individuals to track conversational threads, follow cause-and-effect relationships in real-time problems, and mentally simulate prospective actions before executing them.

3.3 Interface with Long-Term Memory

In the original 1974 multicomponent architecture, the interface between short-term working memory and long-term memory was conceptualized vaguely, primarily routing through the central executive. The 2000 revision established the Episodic Buffer as the primary representational gateway connecting immediate working memory to the permanent repositories of explicit long-term memory (semantic and episodic memory).

This interface operates bidirectionally, supporting both top-down and bottom-up cognitive workflows:

• Top-Down Semantic Interpretation: When sensory stimuli enter working memory, the Episodic Buffer retrieves contextual schemas and semantic frameworks from long-term memory to interpret and organize the incoming stream. For example, during reading, lexical forms and grammatical rules stored in long-term memory are deployed into the Episodic Buffer to assemble words into meaningful propositions.
• Bottom-Up Episodic Consolidation: The Episodic Buffer functions as an initial staging ground for the creation of new episodic memories. By synthesizing multi-sensory features into unified perceptual chunks over several seconds, the buffer prepares integrated representational packets suitable for long-term consolidation.

It is vital to distinguish the short-term maintenance of an episode within the buffer from its enduring structural consolidation via the medial temporal lobe. The Episodic Buffer maintains an episode only across transient operational intervals (seconds to minutes) through active neural firing and synchronized network oscillation. Permanent long-term storage requires synaptic plasticity and systemic consolidation orchestrated by the hippocampus and neocortex over periods of hours, days, and weeks. A patient with damage to the hippocampal consolidation machinery can successfully generate and examine an episode within the buffer, but once their attention shifts, that episode dissipates without leaving an enduring memory trace.

3.4 Neuroanatomical and Experimental Substrates

While the Central Executive is heavily localized within the frontoparietal Central Executive Network, the Episodic Buffer exhibits a more distributed neuroanatomical footprint, reflecting its role as a cross-modal integrator and bridge to long-term memory.

Neuroimaging and lesion studies implicate three primary neural systems in the operations of the Episodic Buffer:

Medial Temporal Lobe (MTL) and Hippocampal Formations: The hippocampus (specifically areas CA1, CA3, and the dentate gyrus) and surrounding parahippocampal, entorhinal, and perirhinal cortices are critically involved in the rapid relational binding of disparate features. While dense amnesics with selective hippocampal lesions retain basic buffer functions for highly practiced, familiar narrative material, novel and arbitrary cross-modal associations (such as linking an unfamiliar face to an abstract geometric shape and an arbitrary auditory tone) show profound degradation when hippocampal structures are damaged.

Posterior Parietal Cortex and the Temporoparietal Junction (TPJ): Neuroimaging paradigms using functional Magnetic Resonance Imaging (fMRI) repeatedly identify the inferior parietal lobule, particularly the angular gyrus (BA 39), as an essential neural hub for multimodal semantic integration. The angular gyrus functions as a convergence zone where visual, auditory, and spatial processing streams intersect, facilitating the binding of multi-sensory representations into coherent mental experiences.

Ventrolateral Prefrontal Cortex (vlPFC): While the dlPFC manages supervisory control, the vlPFC (Brodmann Areas 44, 45, and 47) is heavily implicated in the controlled retrieval of representations from long-term memory and the structural selection of multimodal chunks. The left vlPFC (including Broca’s area) interfaces directly with semantic memory stores, assisting in the linguistic structuring of episodic chunks within the buffer.

At the electrophysiological level, empirical evidence indicates that information binding within the Episodic Buffer is achieved through synchronized oscillatory activity across distributed cortical zones. Phase-amplitude coupling—specifically the nesting of high-frequency gamma oscillations (30–80 Hz) within lower-frequency theta rhythms (4–8 Hz)—provides the timing mechanism for holding and sequencing multiple bound chunks within human neocortex. Magnetoencephalography (MEG) studies demonstrate that when participants successfully maintain bound visual-auditory objects in working memory, cross-cortical coherence surges between auditory temporal regions, visual occipital regions, and the prefrontal cortex, providing clear biological evidence of an integrated buffer workspace.

4. Key Differences Between Central Executive and Episodic Buffer

4.1 Primary Function: Processing Control vs. Multimodal Storage

The distinction between the Central Executive and the Episodic Buffer is grounded in their operational nature: the Central Executive is fundamentally a processor, whereas the Episodic Buffer is fundamentally a store.

The Central Executive possesses no internal mnemonic registers; it does not hold words, numbers, images, or multimodal scenes. Its functions are entirely supervisory. It coordinates cognitive operations by altering the activation levels of neural pathways, steering the direction of perceptual sensors, suppressing task-irrelevant impulses, and allocating computational resources. If cognition is compared to an industrial manufacturing plant, the Central Executive is the plant manager: directing work orders, resolving scheduling conflicts, and prioritizing assembly lines, without physically housing or retaining the manufactured products.

The Episodic Buffer, conversely, is an informational repository. It exists to hold data in an integrated format across short temporal windows. It does not formulate operational goals, execute task-switching routines, or resolve abstract cognitive conflicts. Instead, it provides the representational canvas upon which data from the phonological loop, the visuospatial sketchpad, and long-term memory are temporarily assembled, manipulated, and rendered accessible to consciousness. Continuing the industrial analogy, the Episodic Buffer is the central assembly staging deck: the physical space where parts from different supply lines are temporarily arranged together into a unified, coherent prototype.

4.2 Representational Medium and Code

The two components operate through fundamentally different cognitive mechanisms and representational formats:

The Central Executive operates via domain-general control signals. It does not utilize an informational code in the traditional sense; rather, it produces top-down neuromodulatory biases. It signals what to prioritize and what to suppress across distinct cognitive systems. Its operations are amodal and procedural, lacking qualitative sensory features, semantic dimensions, or spatial geometries.

The Episodic Buffer utilizes a multidimensional representational code. This represents a distinct form of cognitive information: a standardized, abstract format capable of encoding acoustic frequencies, phonemic categories, spatial locations, chromatic values, morphological structures, and deep semantic meanings simultaneously. Through this multidimensional code, a bound chunk inside the episodic buffer is not merely a collection of isolated features; it is an integrated token. When an individual recalls a scene from an autobiographical event, the buffer holds the visual color of an object, the voice of the speaker, the physical layout of the room, and the emotional context as an indivisible cognitive representation.

4.3 Attentional Dependency and Conscious Awareness

One of the most complex theoretical debates concerning the relationship between these two systems centers on attentional dependency: does the Episodic Buffer require continuous attention from the Central Executive to bind and maintain its multimodal representations?

In Baddeley’s initial 2000 formulation, the Episodic Buffer was conceptualized as highly dependent on the Central Executive. It was hypothesized that the process of cross-modal binding required continuous executive attentional resources. Under this model, binding was an active, computationally expensive executive operation; without sustained attention, bound representations would immediately dissolve into their component sensory features.

However, subsequent experimental research overturned this assumption. In a series of empirical investigations using dual-task paradigms, Alan Baddeley, Graham Hitch, and colleagues examined cross-modal visual-verbal binding under heavy executive cognitive load. Participants were presented with shapes associated with arbitrary spoken names while simultaneously performing demanding executive distraction tasks (such as backward counting or random sequence generation). The results demonstrated that while the secondary executive task degraded overall performance, it disrupted the retention of bound combinations no more severely than it disrupted the retention of individual, isolated features. Feature binding itself proved to be largely automatic.

These findings led to the revised model (Baddeley, 2012), which established that initial cross-modal feature integration occurs pre-attentively and automatically before delivery to the Episodic Buffer. The buffer can passively hold pre-bound multidimensional chunks without requiring continuous executive intervention. The role of the Central Executive is not to glue the features together, but rather to direct the beam of focal attention onto the buffer’s contents, manipulate those chunks to solve problems, and selectively retrieve specific items into conscious awareness.

Concurrently, the Episodic Buffer is recognized as the primary representational platform for conscious phenomenological experience (qualia) within working memory. While the Central Executive directs the spotlight of attention, the Episodic Buffer is the illuminated stage. Conscious awareness requires a coherent, multimodal, temporally organized representation of current reality; the Episodic Buffer provides the cognitive workspace where that multidimensional model is actively maintained.

4.4 Comparison Dimensions Summary

To systematically summarize the structural and functional dissociations between the Central Executive and the Episodic Buffer, the head-to-head comparison below presents their specific cognitive parameters, theoretical boundaries, and neuropsychological profiles.

Functional Dimension	Central Executive (CE)	Episodic Buffer (EB)
Architectural Role	Supervisory attentional controller; manages resource allocation and task scheduling.	Multimodal storage reservoir; integrates cross-system outputs into coherent episodes.
Mnemonic Storage Capacity	Zero internal storage; acts solely as a processing and gating mechanism.	Limited capacity; approximately 4 multidimensional chunks (Cowan’s limit).
Representational Medium	Amodal, domain-general top-down attentional and neuromodulatory control signals.	Multidimensional code capable of expressing visual, spatial, acoustic, and semantic features.
Primary Operations	Inhibition, set-shifting, task-coordination, selective and divided attention.	Cross-modal binding, temporary multimodal maintenance, chronological sequencing.
Interface with Long-Term Memory (LTM)	Regulates retrieval strategies, guides memory search, and suppresses irrelevant memories.	Serves as the structural workspace where retrieved LTM semantic schemas bind with active perceptual inputs.
Phenomenological Status	Directs the spotlight of conscious focus; does not itself embody qualitative conscious states.	Provides the representational platform for conscious awareness; houses the contents of current qualia.
Neuroanatomical Substrates	Frontoparietal network: Dorsolateral Prefrontal Cortex (dlPFC), Anterior Cingulate Cortex (ACC), Intraparietal Sulcus.	Temporoparietal junction (angular gyrus), Medial Temporal Lobes (hippocampus), Ventrolateral Prefrontal Cortex (vlPFC).
Clinical Pathology Profile	Dysexecutive Syndrome: Perseveration, distractibility, disinhibition, loss of cognitive flexibility.	Associative Binding Deficits: Inability to hold complex prose passages or novel multimodal combinations across brief delays.
Primary Diagnostic Paradigms	Stroop Task, Wisconsin Card Sorting Test (WCST), N-Back (updating), Task-Switching Paradigms.	Immediate Sentence/Prose Recall, Cross-Modal Feature Binding Tasks, Complex Working Memory Span Tasks.

The comparative analysis demonstrates that the Central Executive and the Episodic Buffer possess complementary properties. The Central Executive maximizes dynamic operational flexibility at the cost of zero representational permanence, while the Episodic Buffer provides integrated representational stability within a four-chunk window, operating as a passive or semi-passive workspace for executive manipulation.

5. Similarities and Functional Overlap

5.1 Shared Contributions to Complex Cognition

While theoretical modeling requires rigorous conceptual dissociation between the Central Executive and the Episodic Buffer, high-level cognition relies on their seamless collaboration. In real-world environments, these two systems operate as a tightly coupled processing unit, forming the cognitive engine of human fluid intelligence (Gf).

Both systems represent the “central core” of Baddeley’s model, clearly distinguished from the peripheral slave systems (the phonological loop and the visuospatial sketchpad). The peripheral loops are restricted to unimodal sensory maintenance; they are peripheral conduits that isolate acoustic or spatial details. In contrast, both the Central Executive and the Episodic Buffer operate across sensory domains. Neither is restricted to a single perceptual stream; both manage abstract, complex, and cross-modal operations.

Furthermore, both systems are primary contributors to performance on complex span tasks (such as the Reading Span, Operation Span, and Symmetry Span tasks developed by Daneman, Carpenter, Engle, and colleagues). Complex span tasks require an individual to alternate rapidly between processing a distracting cognitive problem (e.g., verifying a mathematical equation or validating a sentence’s grammar) and maintaining a sequence of target items (e.g., words or spatial locations) across a delayed testing interval. Extensive psychometric literature reveals that scores on complex span tasks predict real-world cognitive performance—including reading comprehension scores, standardized academic achievements, complex decision-making proficiency, and computer programming aptitude—far better than simple span tasks (like forward digit span). This high predictive validity occurs because complex span tasks load concurrently on both the Central Executive (which manages task-switching, processing throughput, and inhibition of proactive interference) and the Episodic Buffer (which holds the target items in a multidimensional, bound state despite continuous distraction).

5.2 Interdependent Dynamic Operational Loop

In practice, the Central Executive and the Episodic Buffer function in a continuous, dynamic feedback loop. This relationship can be conceptualized as an ongoing cycle of attentional selection, representational stabilization, evaluation, and strategic updating.

The Central Executive exercises selective top-down control over what informational streams are permitted to cross the threshold into the Episodic Buffer. Through its attentional gating mechanisms, the executive selects specific perceptual outputs from the peripheral slave systems and coordinates the targeted retrieval of relevant schemas from long-term memory. Without this executive gating, the limited four-chunk capacity of the Episodic Buffer would instantly be overwhelmed by irrelevant environmental noise and competing associative memories.

Once bound chunks are instantiated within the Episodic Buffer, the buffer feeds this integrated informational model back to the Central Executive. The executive inspects the multidimensional representation, evaluates its properties against current behavioral goals, and makes strategic determinations. If an unexpected obstacle or conceptual ambiguity is detected within the buffer’s episodic model (such as a plot twist in a story or an unexpected calculation error in a math problem), the Central Executive reconfigures its attentional priorities, shifts cognitive sets, and directs the retrieval of alternative schemas to resolve the discrepancy.

Because of this high functional interdependency, both structures exhibit simultaneous vulnerability to cognitive strain and physiological degradation:

• Cognitive Fatigue and Overload: Under conditions of sleep deprivation, high environmental stress, or prolonged mental effort, both executive control and buffer capacity degrade in tandem. Attentional filters weaken, permitting task-irrelevant distractors to enter the buffer, while the capacity to hold integrated, multi-feature chunks without informational decay declines.
• Neurodegenerative Disease: In early-stage Alzheimer’s disease, pathological amyloid-beta aggregation and neurofibrillary tau tangles accumulate in the transentorhinal regions, hippocampus, and temporoparietal areas, before spreading to the frontal neocortex. Consequently, patients exhibit a progressive breakdown of both episodic buffering (manifesting as profound failures to bind recent cross-modal details) and executive control (manifesting as perseveration, impaired goal maintenance, and poor attentional regulation).

6. Cognitive Application and Functional Roles Across Tasks

6.1 Narrative and Prose Comprehension

The real-time comprehension of written or spoken language illustrates how the Central Executive and the Episodic Buffer divide their computational labor during continuous cognitive tasks.

As a reader scans a complex narrative text, linguistic decoding begins with early visual processing of orthographic symbols, which are rapidly converted into subvocal phonological sequences managed by the phonological loop. However, understanding a story requires far more than retaining a series of phonetic sounds. This is where the Episodic Buffer takes over representational duty. The buffer continuously constructs, maintains, and updates a situation model—a coherent mental representation of the characters, spatial settings, actions, temporal trajectories, and psychological motivations described in the narrative.

To build this situation model, the Episodic Buffer draws upon the reader’s long-term semantic and episodic memory to populate the scene with world knowledge. If the text mentions that a character “entered a café and ordered an espresso,” the buffer integrates the lexical tokens with the long-term semantic schema of what a café looks like, the sequence of paying a cashier, and the sensory qualia of coffee. The buffer retains the immediate chronological sequence of events, ensuring that the reader understands who spoke to whom, which action preceded another, and how individual sentences link logically to one another.

Simultaneously, the Central Executive operates as the supervisory regulator of this comprehension process. Its functional roles during reading include:

• Lexical Ambiguity Resolution: When encountering homonyms or polysemous words (e.g., “bank” referring to a financial institution versus a riverbank), the Central Executive suppresses the context-inappropriate interpretation, allowing only the task-relevant meaning to integrate into the buffer’s situation model.
• Goal-Directed Attentional Shifting: If a paragraph presents complex scientific reasoning or dense character interactions, the Central Executive slows the pace of reading, commands saccadic eye regressions to re-read previous sentences, and shifts attention between low-level lexical decoding and high-level thematic inference.
• Suppression of Irrelevant Associations: The executive ensures that idiosyncratic personal memories triggered by the text do not intrude into conscious awareness and displace the four-chunk working memory capacity of the buffer.

If an individual possesses a robust Central Executive but an impaired Episodic Buffer, they can parse individual grammatical rules and maintain isolated words, but they will completely lose the narrative arc of a paragraph, failing to synthesize multiple propositions into a coherent global scenario. Conversely, if an individual possesses an intact Episodic Buffer but an impaired Central Executive, they may retain immediate story chunks but will struggle to filter out distracting inferences, reconcile plot ambiguities, or adjust their reading strategy when contextual cues shift.

6.2 Complex Problem-Solving and Mathematics

Mental mathematics and structured problem-solving provide clear examples of how executive processing operates on multi-feature working memory representations.

Consider the cognitive operations required to perform multi-step mental arithmetic, such as calculating the product of 47 multiplied by 36 without external writing aids:

1. The Central Executive decomposes the global calculation into algorithmic sub-problems: $(40 \times 30) + (40 \times 6) + (7 \times 30) + (7 \times 6)$.
2. The phonological loop and visuospatial sketchpad manage immediate numerical digits and spatial column layouts.
3. The Episodic Buffer acts as the central scratchpad, temporarily holding the intermediate mathematical products. It preserves the intermediate solution of $1200$, integrates the next solution of $240$ into a running total of $1440$, and maintains this chunk alongside the spatial representation of the problem and the verbal instructions regarding which algorithmic step comes next.
4. The Central Executive directs the computational transitions, constantly monitoring progress, scheduling the execution of arithmetic rules retrieved from long-term memory, and inhibiting intrusive numbers from previous calculation steps.

Without the Episodic Buffer, the intermediate sums would vanish as soon as the executive diverted its attentional resources toward executing the next multiplication step. Without the Central Executive, the intermediate numbers held within the buffer would remain static; there would be no operational engine to manipulate them, calculate their sum, or verify the accuracy of the result.

6.3 Clinical Dissociations and Neuropsychological Profiles

In neuropsychology, the validity of a theoretical distinction depends on demonstrating double dissociations—identifying patients in whom Process A is impaired while Process B is preserved, alongside patients in whom Process B is impaired while Process A remains intact. The distinction between the Central Executive and the Episodic Buffer is supported by clear clinical evidence across diverse patient populations.

Profile 1: Intact Episodic Buffer with Impaired Central Executive (Frontal Lobe Lesions). Patients suffering from focal lesions to the dorsolateral prefrontal cortex or anterior cingulate cortex—often resulting from traumatic brain injury, stroke, or behavioral variant frontotemporal dementia—display classic dysexecutive syndrome. When administered executive testing batteries (e.g., the Stroop Task, the Wisconsin Card Sorting Test, the Trail Making Test Part B), these individuals exhibit severe impairments: perseverative responding, an inability to inhibit prepotent impulses, and poor dual-task coordination.

However, when presented with rich, structured narrative material that does not require active rule switching or heavy inhibitory filtering, their Episodic Buffer capacity often remains intact. In immediate prose recall tests, these patients can accurately recall immediate stories of twelve to sixteen words without significant degradation, because their multimodal binding mechanisms and temporoparietal representational networks are structurally undamaged. They can temporarily hold a bound, coherent episode as long as the testing environment is structured to prevent distracting sensory inputs from triggering disinhibited behavioral routines.

Profile 2: Intact Central Executive with Impaired Episodic Buffering (Early Medial Temporal/Temporoparietal Pathology). Conversely, in early stages of specific neurodegenerative conditions, such as atypical presentations of Alzheimer’s disease or semantic dementia variants involving temporoparietal junction degradation, patients present with an inverted cognitive profile. These individuals perform well on tests of pure attentional allocation, rule switching, and inhibitory control. They can suppress prepotent responses on a Stroop task and transition between sorting rules on the WCST with normal accuracy.

Yet, when administered tests requiring the immediate cross-modal binding of arbitrary features—such as recalling which specific, unfamiliar abstract shape was paired with which specific non-word name—their performance collapses. They cannot hold more than one or two multi-feature chunks simultaneously, and their immediate recall of connected prose drops to levels seen with unrelated word lists. Their Central Executive functions normally, but the multimodal representational workspace of the Episodic Buffer is degraded, preventing them from synthesizing disparate sensory inputs into cohesive working memory representations.

These clinical dissociations carry immediate implications for neuropsychological assessment and cognitive rehabilitation, as summarized below:

• When to assess the Central Executive: Evaluations should focus on executive control when the patient demonstrates real-world organizational failures, poor impulse control, disinhibition, perseverative errors, high distractibility, or an inability to navigate changes in routine. Clinicians deploy the Stroop task, Trail Making Test Part B, Tower of London, and N-back paradigms to quantify processing regulation and attentional flexibility.
• When to assess the Episodic Buffer: Evaluations should focus on episodic buffering when the patient presents with complaints of losing track of multi-step verbal-visual instructions, difficulty synthesizing information across sensory channels, or marked differences between immediate memory for isolated words versus structured sentences. Clinicians deploy immediate narrative recall subtests (e.g., the Logical Memory subtest of the Wechsler Memory Scale) and cross-modal associative binding tasks.
• Rehabilitation Strategy Dissociation: For executive deficits, cognitive rehabilitation emphasizes external pacing, environmental modifications to eliminate distractions, checklist cues, and meta-cognitive strategy training (e.g., Goal Management Training). For episodic buffer deficits, rehabilitation focuses on reducing information density, breaking instructions down into single-modality steps (to prevent cross-modal binding overloads), and utilizing externalized associative memory prostheses.

7. Frequently Asked Questions

7.1 Why was the Episodic Buffer added 26 years after the original model?

The addition of the Episodic Buffer to Baddeley’s multicomponent working memory model in the year 2000 occurred after decades of accumulating empirical contradictions that the original 1974 tripartite framework could not reconcile. The original model was designed around three components: the phonological loop, the visuospatial sketchpad, and the central executive. While this architecture successfully accounted for simple dual-task performance involving isolated acoustic or visual tasks, it possessed three major structural limitations:

1. Prose and Sentence Memory: The model could not explain why immediate memory span for words arranged into meaningful sentences (15–20 words) is vastly larger than the span for unrelated word lists (5–6 words). The phonological loop lacked the physical storage capacity to hold this linguistic material, and the central executive explicitly lacked storage capacity.
2. Preserved Narrative Memory in Dense Amnesia: Research on patients with bilateral medial temporal lobe damage revealed that densely amnesic patients—who were incapable of encoding memories into long-term storage—could nonetheless recall complex, multi-sentence narrative stories accurately over short delays. This proved that immediate episodic representation did not depend on long-term memory consolidation.
3. Absence of a Cross-Modal Binding Mechanism: Real-world experience requires the continuous integration of visual, acoustic, spatial, and semantic data. The original model contained no direct informational bridge between the phonological loop and the visuospatial sketchpad.

Baddeley introduced the Episodic Buffer to resolve these anomalies without dismantling the core principles of the original model. The update provided a dedicated, limited-capacity multimodal store that bridged the specialized peripheral loops, the central executive, and long-term memory.

7.2 Does the Central Executive possess any storage capacity?

Under Alan Baddeley’s established multicomponent framework—as formulated following his 1986 adoption of the Norman-Shallice Supervisory Attentional System and cemented in the 2000 update—the Central Executive possesses strictly zero mnemonic storage capacity. It is exclusively an attentional processor, coordinator, and gating engine.

All informational retention within the multicomponent architecture is handled by the three storage systems: the Phonological Loop (acoustic and verbal storage), the Visuospatial Sketchpad (visual and spatial storage), and the Episodic Buffer (multimodal, integrated chunk storage). Any working memory capacity historically attributed to the Central Executive represented unisolated activity within the Episodic Buffer. When an individual maintains numbers, facts, images, or concepts during a problem-solving sequence, those representations reside inside the peripheral buffers or the Episodic Buffer, while the Central Executive regulates attention and manipulates those traces.

It is important to note that alternative working memory frameworks—such as Nelson Cowan’s Embedded-Processes Model or Randall Engle’s Executive Attention Theory—characterize working memory capacity differently. Cowan defines working memory capacity as the amount of information within the “focus of attention” (approximately four chunks), viewing executive control and immediate storage as integrated aspects of activated long-term memory. However, within Baddeley’s multicomponent taxonomy, the distinction remains firm: the Central Executive provides attentional processing without storage, while the Episodic Buffer provides multimodal storage without supervisory control.

7.3 How does the Episodic Buffer interact with Long-Term Episodic Memory?

The Episodic Buffer serves as an active, temporary workspace that interfaces directly with permanent, long-term episodic memory, but the two systems are dissociable across temporal, neuroanatomical, and operational dimensions:

• Temporal Duration: The Episodic Buffer maintains representations across a transient window of seconds to a few minutes, relying upon active neural firing and synchronized oscillations. Long-term episodic memory preserves experiences across hours, weeks, years, and decades through enduring structural and synaptic modifications (such as long-term potentiation).
• Direction of Information Flow: During cognitive processing, the Episodic Buffer retrieves established schemas, semantic contexts, and past autobiographical details from long-term memory to interpret and structure incoming perceptual experiences. In return, the Episodic Buffer binds novel, multi-sensory perceptual combinations into integrated packets, which are then passed to the medial temporal lobe system for long-term consolidation.
• Neuroanatomical Dissociation: Long-term episodic consolidation depends fundamentally on the integrity of the hippocampus and parahippocampal structures. If these structures undergo bilateral destruction, the patient can no longer consolidate new permanent memories. However, as demonstrated in amnesic patients, the short-term maintenance of an integrated narrative within the Episodic Buffer remains intact, supported by temporoparietal convergence zones and frontoparietal networks.

7.4 Can the Episodic Buffer function without the Central Executive?

The degree of functional independence between the Episodic Buffer and the Central Executive depends on whether one is examining the initial binding of information or the deliberate manipulation of those bound representations.

Empirical evidence demonstrates that the initial cross-modal binding of sensory features can occur automatically and pre-attentively, without requiring the focal attention of the Central Executive. In dual-task experiments, forcing participants to perform demanding executive tasks (such as continuous random number generation) does not disproportionately disrupt their ability to bind shapes with colors or spoken words with visual symbols. The sensory and semantic features integrate into unified multidimensional chunks via bottom-up mechanisms, entering the Episodic Buffer automatically.

However, sustaining, manipulating, and strategically updating those representations requires the Central Executive. Once bound representations reside within the buffer, an individual cannot inspect them, update their components, use them to guide multi-step reasoning, or shield them from proactive interference without executive attention. Patients with severe executive impairments can maintain a passive bound chunk across brief delays if they are kept free from environmental distraction; however, as soon as novel stimuli appear, the absence of executive gating results in immediate memory disruption.

8. Conclusion

8.1 Synthesis of Cognitive Contributions

The fractionation of working memory into specialized processing and storage components represents a foundational achievement in cognitive psychology. Within this structural architecture, the Central Executive and the Episodic Buffer resolve the tension between dynamic computational control and integrated mental representation.

The Central Executive operates as the supervisory engine of cognition. It is an attentional coordinator that allocates resources, resolves competing demands, shifts task sets, and inhibits inappropriate behavioral impulses. Operating with zero internal storage capacity, it provides the flexible, goal-directed control that prevents human behavior from degenerating into a series of automated, reflexive responses to immediate environmental stimuli.

The Episodic Buffer serves as the multimodal canvas of cognition. It provides a limited-capacity, four-chunk storage space that translates sensory-specific codes into an integrated, multidimensional format. By binding the acoustic data of the phonological loop, the spatial configurations of the visuospatial sketchpad, and the semantic schemas of long-term memory into an organized chronological sequence, the buffer establishes the representational workspace necessary for human conscious awareness.

Together, these two components explain how the human mind can continuously interact with a dynamic, multifaceted sensory environment. The Central Executive directs the spotlight of processing capacity, while the Episodic Buffer holds the illuminated, integrated representations. This division between processing regulation and multimodal storage provides a cohesive framework for understanding both normal cognitive operations and diverse clinical neurocognitive disorders.

8.2 Future Directions in Working Memory Research

As cognitive neuroscience advances, the theoretical frameworks governing the Central Executive and the Episodic Buffer continue to evolve through several active lines of investigation:

High-Resolution Dynamic Connectomics: Utilizing ultra-high-field functional neuroimaging (7-Tesla fMRI) and magnetoencephalography (MEG), researchers are shifting from modular, static localization toward dynamic functional connectomics. Investigations focus on tracking how the frontoparietal Central Executive Network dynamically couples and decouples from the default mode network and posterior parietal integration hubs (such as the angular gyrus) on a millisecond-by-millisecond basis during complex task execution.

Computational Modeling of Neural Oscillations: Computational neuroscientists are constructing biophysical models of information binding within the Episodic Buffer. Research centers on how cross-frequency coupling—specifically theta-gamma phase-amplitude coupling across distributed sensory and associative cortices—provides a biological mechanism for multiplexing distinct, bound informational chunks without cross-talk or catastrophic interference.

Refined Clinical Biomarkers for Neurodegenerative Diseases: Clinical neuroscientists are working to isolate fine-grained behavioral assays that differentiate early executive control degradation from episodic binding failures. Developing sensitive computerized tasks that specifically target cross-modal feature binding offers substantial promise for identifying preclinical stages of Alzheimer’s disease before overt structural atrophy or irreversible cognitive decline occurs.

Theoretical Reconciliation Between Multicomponent and State-Based Models: The field continues to debate the relationship between Baddeley’s structural multicomponent architecture and state-based, emergent-process models (such as those of Cowan, Engle, and Oberauer). Future empirical work will likely focus on unifying these perspectives, establishing whether the Episodic Buffer is best conceptualized as a distinct neuroanatomical storage chamber or as an emergent state of activated long-term memory held within the primary focus of executive attention.

References

• Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory: A proposed system and its control processes. In K. W. Spence & J. T. Spence (Eds.), The Psychology of Learning and Motivation (Vol. 2, pp. 89–195). Academic Press. https://doi.org/10.1016/S0079-7421(08)60422-3
• Baddeley, A. D. (1986). Working Memory. Oxford University Press.
• Baddeley, A. (1996). Exploring the central executive. The Quarterly Journal of Experimental Psychology Section A, 49(1), 5–28. https://doi.org/10.1080/713755608
• Baddeley, A. (2000). The episodic buffer: A new component of working memory? Trends in Cognitive Sciences, 4(11), 417–423. https://doi.org/10.1016/S1364-6613(00)01538-2
• Baddeley, A. (2012). Working memory: Theories, models, and controversies. Annual Review of Psychology, 63, 1–29. https://doi.org/10.1146/annurev-psych-120710-100422
• Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. H. Bower (Ed.), The Psychology of Learning and Motivation (Vol. 8, pp. 47–89). Academic Press. https://doi.org/10.1016/S0079-7421(08)60452-1
• Baddeley, A. D., & Wilson, B. A. (2002). Prose recall and amnesia: Implications for the working memory model. Neuropsychologia, 40(10), 1737–1743. https://doi.org/10.1016/S0028-3932(01)00146-4
• Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87–114. https://doi.org/10.1017/s0140525x01003922
• Daneman, M., & Carpenter, P. A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior, 19(4), 450–466. https://doi.org/10.1016/S0022-5371(80)90312-6
• Engle, R. W. (2002). Working memory capacity as executive attention. Current Directions in Psychological Science, 11(1), 19–23. https://doi.org/10.1111/1467-8721.00160
• Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex “frontal lobe” tasks: A latent variable analysis. Cognitive Psychology, 41(1), 49–100. https://doi.org/10.1006/cogp.1999.0734
• Norman, D. A., & Shallice, T. (1986). Attention to action: Willed and automatic control of behavior. In R. J. Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and Self-Regulation (pp. 1–18). Springer. https://doi.org/10.1007/978-1-4684-4778-1_1