Context-Dependent Memory vs. State-Dependent Memory

The human memory system does not operate like an immutable digital archive where discrete packets of data are cataloged in isolation and retrieved untouched on demand. Rather, the architecture of human cognition is intrinsically dynamic, reconstructive, and fundamentally dependent upon the presence of informational prompts known as retrieval cues. For decades, cognitive psychologists and neurobiologists have recognized that an individual’s capacity to access stored mnemonic representations is governed not merely by the structural integrity of the original memory trace, but by the precise relationship between the conditions present during information encoding and those present during the attempt at retrieval. When environmental features or internal somatic parameters change markedly between learning and testing, recall performance routinely falters, giving rise to the pervasive cognitive phenomenon known as cue-dependent forgetting.

Within this theoretical landscape, two distinct yet frequently conflated manifestations of cue-dependent retrieval dominate the experimental literature: context-dependent memory and state-dependent memory. While both paradigms demonstrate that human recall is profoundly sensitive to situational congruence, they diverge along critical biological, sensory, and functional axes. Context-dependent memory highlights the influence of the external, physical surroundings—the ambient sights, sounds, smells, and spatial geometries of the objective world. Conversely, state-dependent memory centers upon the organism’s internal milieu—the neurochemical, hormonal, physiological, and affective configurations that define subjective somatic reality at any given moment. Disentangling the mechanisms governing these two modes of retrieval is essential for constructing comprehensive models of human cognition and for applying memory science to real-world educational, clinical, and forensic environments.

This treatise provides an exhaustive, comparative analysis of context-dependent and state-dependent memory. By examining their shared roots in associative network theory and the encoding specificity principle, delineating their divergent neurobiological substrates, and evaluating their operational parameters across laboratory and applied domains, this article establishes a definitive framework for understanding how physical settings and biological states dictate the boundaries of human recollection.

1. Introduction to Retrieval Cue Phenomena

1.1 Defining Context-Dependent and State-Dependent Memory

Context-dependent memory refers to the empirical phenomenon whereby an individual demonstrates superior retention and recall of information when the external environmental conditions at the time of retrieval match those that were present during the initial encoding phase. In this framework, the ambient stimuli comprising the learner’s surroundings—such as room geometry, lighting conditions, background acoustics, atmospheric pressure, and even ambient olfactory signatures—are encoded alongside the target material. These peripheral, incidental features become integrated into the episodic memory trace. Consequently, re-exposure to these specific physical cues during retrieval facilitates access to the target representation by serving as navigational conduits within semantic and episodic neural networks.

In contrast, state-dependent memory describes an enhancement in memory performance that occurs when an individual’s internal physiological, neurochemical, or psychological condition at the time of retrieval corresponds to their state during information acquisition. Rather than relying on external environmental cues, state dependency is governed by interoceptive variables, including autonomic nervous system activation, drug-induced neurochemical alterations, hormonal levels, somatic exhaustion, and endogenous affective states. When an organism learns material while under the influence of a specific physiological state—such as elevated plasma cortisol or stimulant-induced tachycardia—the interoceptive sensations and neurochemical pathways active during that state become bound to the memory trace. Retrieval attempts conducted in an altered or mismatched physiological state often suffer from marked decrements, as the necessary internal cues are absent.

Both of these phenomena reside under the theoretical umbrella of cue-dependent forgetting and retrieval failure theories. According to these frameworks, the inability to recall a previously consolidated piece of information does not inevitably imply trace decay or structural erasure from the brain; rather, it reflects a temporary failure of accessibility attributable to the absence of appropriate retrieval cues. Clarifying the precise boundary between external environmental determinants and internal somatic determinants is critical for cognitive neuroscience, contemporary educational engineering, psychotherapeutic interventions for trauma, and legal evaluations of eyewitness testimony.

1.2 Historical Foundations and the Encoding Specificity Principle

Historically, early 20th-century memory paradigms treated the human mnemonic system as an essentially static repository. Grounded in classical associationism and early behaviorist models, memory retention was conceptualized primarily as a function of trace strength, rehearsal frequency, and the passage of time, as formalized by the traditional Ebbinghaus forgetting curve. In these models, forgetting was largely attributed to passive decay or proactive and retroactive interference caused by competing informational inputs. The contextual background against which learning occurred was dismissed as sensory noise, secondary to the focal target stimulus.

This mechanistic, storage-centric paradigm shifted dramatically in 1973 when cognitive psychologists Endel Tulving and Donald M. Thomson formulated the Encoding Specificity Principle. Tulving and Thomson posited that no memory cue, no matter how strongly associated it might be semantically with a target item in normative language, is universally effective. Instead, a retrieval cue is effective only to the extent that its informational properties were specifically integrated into the memory trace during the initial encoding event. The Encoding Specificity Principle posited that episodic memory traces are holistic, configural records composed of both the focal item and the multivariant background features present during acquisition. Memory performance is therefore fundamentally an interactive function of the informational overlap between the stored representation and the cues present at the retrieval interface.

The establishment of this principle sparked a wave of environmental and physiological manipulation studies in the late 1970s and 1980s. Researchers began deliberately altering external classrooms, underwater dive sites, and pharmacological states to test the boundaries of retrieval cue overlap. In modern cognitive neuroscience, this construct has expanded across synaptic, systemic, and behavioral dimensions. Contemporary models demonstrate that the brain does not store memories in isolated cortical silos; rather, it coordinates dynamic, widely distributed ensembles of neurons across the hippocampus, parahippocampal cortex, and prefrontal networks that inherently encode ambient physical contexts and interoceptive somatic states as indivisible components of the mnemonic experience.

1.3 Why the Concepts Are Frequently Conflated

Despite their distinct conceptual and neurobiological boundaries, context-dependent memory and state-dependent memory are frequently conflated within both academic discourse and general pedagogical literature. A primary cause of this confusion is their shared mathematical and structural representation within associative network models of memory. In computational cognitive architectures, such as John R. Anderson’s ACT-R or spreading activation frameworks, both external environmental stimuli and internal somatic markers are modeled identically: as context nodes linked via associative vectors to target proposition nodes. Because the algorithmic mechanics of activation spread from a “room node” or an “arousal node” to a “target item node” are mathematically indistinguishable in computational simulations, observers often collapse them into a single, generic category of “contextual cueing.”

A second, more profound source of conflation arises from the natural co-occurrence and bidirectional coupling of physical environments and internal physiological states in ecological settings. Under naturalistic conditions, an organism’s external surroundings routinely trigger predictable autonomic, hormonal, and affective shifts. Entering a sterile, harshly lit hospital wing or an echoic, high-ceilinged examination hall rarely constitutes an isolated change in visual or auditory context; rather, these specific physical environments immediately elicit heightened sympathetic nervous system activity, surges in epinephrine, and elevated subjective anxiety. When an individual struggles to retrieve information in such an environment, the retrieval failure is often driven by a compound misalignment of both exogenous and endogenous cues, rendering precise attribution difficult without specialized experimental intervention.

Furthermore, introductory textbooks and popular psychological writing frequently dilute these concepts, using terms such as “situational memory,” “environmental dependency,” and “state dependency” interchangeably to describe any instance of cue-dependent retrieval. Disentangling external environmental inputs from internal neurochemical variables requires rigorous experimental designs—such as keeping physical settings constant while pharmacologically manipulating autonomic arousal, or holding internal physiological metrics stable across drastically altered physical landscapes. Absent these stringent methodological controls, the boundaries between the world outside the skin and the world within it become readily blurred in casual cognitive discourse.

2. Comprehensive Overview of Context-Dependent Memory

2.1 Theoretical Framework and External Environmental Cues

Context-dependent memory operates on the premise that human cognitive processing is continuously immersed in an exogenous sensory stream that is automatically cataloged alongside focal attention targets. The theoretical architecture of this phenomenon centers entirely on external stimuli that exist independently of the observer’s physiological state. When an individual focuses conscious attention on an explicit learning task—such as memorizing a series of linguistic propositions, mastering motor procedures, or analyzing technical diagrams—the sensory apparati do not restrict their intake exclusively to the target stimuli. Instead, the surrounding environmental array is passively recorded through incidental encoding.

This incidental environmental information encompasses an expansive suite of ambient variables. It includes the spatial configuration and architectural dimensions of the room, the spectral composition and lux levels of the lighting, the acoustic profile of background white noise or mechanical hums, ambient olfactory compounds circulating in the air, and even the local barometric or thermal parameters. Under episodic memory network theories, these diverse exogenous inputs are processed via the primary and secondary sensory cortices and integrated within the medial temporal lobe, specifically the parahippocampal cortex and hippocampus. Rather than filtering out peripheral data as irrelevant sensory noise, the episodic encoding mechanism binds these ambient features directly to the target memory nodes, creating an extensive, interconnected memory web.

When retrieval is attempted, the physical presence of these environmental cues functions as a distributed spatial key. According to pattern completion principles, re-exposure to the peripheral physical features that were present during encoding reactivates the parahippocampal networks that were engaged during learning. This localized reactivation propagates through hippocampal CA3 recurrent collateral connections, facilitating the rapid reconstructive retrieval of the focal target trace. Conversely, when retrieval occurs in a novel or mismatched physical context, the cognitive architecture encounters an absence of these environmental triggers, forcing the individual to rely entirely on internal, deliberate search mechanisms that often fall short of the activation threshold required for successful recall.

2.2 Landmark Experimental Paradigms and Empirical Validation

The definitive empirical validation of context-dependent memory emerged through landmark psychological investigations designed to rigorously test environmental boundary conditions. The most famous of these is the classic 1975 study conducted by D.R. Godden and Alan D. Baddeley, who evaluated scuba divers tasked with learning word lists in two radically contrasting physical environments: on dry land and underwater at a depth of approximately twenty feet. Using a fully crossed 2×2 factorial design (learn dry/test dry, learn dry/test wet, learn wet/test dry, and learn wet/test wet), Godden and Baddeley demonstrated that participants who learned and recalled in matching environments exhibited roughly 40% higher recall accuracy than those whose learning and testing environments diverged. Divergence across this extreme physical threshold introduced profound retrieval decrements, establishing that radical shifts in exogenous sensory context actively inhibit memory access.

Recognizing that the extreme differences between dry land and underwater diving might limit the generalizability of these findings to typical human activities, Steven M. Smith, Arthur Glenberg, and Robert A. Bjork (1978) sought to evaluate environmental context effects within ordinary educational spaces. In a series of meticulously controlled experiments, they had undergraduate students learn lists of paired associates in distinctly different campus rooms. One room was a windowless, cluttered research laboratory tucked away in a basement and equipped with specialized audio equipment; the other was a well-lit, spacious executive conference room on an upper floor overlooking a courtyard. Despite the relatively benign nature of this environmental shift, participants tested in the room identical to their learning environment demonstrated statistically significant recall advantages over those moved to the alternative room. This work confirmed that context dependency was not an artifact of extreme physical stress or survival environments, but an everyday cognitive mechanism.

Subsequent research expanded into alternative sensory domains, notably olfactory contextual cueing. Studies investigating ambient environmental scents have demonstrated that exposing learners to an unfamiliar, distinct olfactory cue (such as novel botanical extracts) during acquisition and subsequently re-introducing that exact airborne scent profile during retrieval significantly enhances verbal recall. Contemporary investigations have expanded these paradigms into digital environments using immersive virtual reality (VR). By utilizing head-mounted displays, modern cognitive researchers can systematically alter fine-grained spatial parameters, visual textures, and architectural dimensions with sub-millimeter precision. These VR studies show that even entirely simulated environmental contexts exert powerful cue-dependent effects on human episodic recall, underscoring that human pattern completion mechanisms react predictably to perceived spatial surroundings regardless of their physical or digital origin.

2.3 Sensory Modalities and Contextual Dimensions

The contextual envelope that encloses an episodic learning experience is multimodal, relying on the convergence of diverse sensory streams to construct an external mental landscape. The primary modalities and structural dimensions of environmental context include:

• Visual Context and Spatial Geometry: The visual configuration of the environment is typically the most salient contextual factor for human learners. This includes the spatial distribution of boundaries, ceiling height, field-of-view depth, interior architectural layout, surface textures, and color palettes. The parahippocampal place area (PPA) and retrosplenial cortex continuously map these structural parameters, creating a holistic spatial frame that anchors the items experienced within it. Changes in focal point distribution or visual field density can disrupt this framing effect.
• Auditory Context and Ambient Acoustics: The ambient auditory landscape exerts consistent, measurable effects on retrieval. This modality includes not only deliberate acoustic backgrounds, such as instrumental music genres, but also non-volitional background variables: the presence of low-frequency ventilation hums, reverberation times determined by room acoustics, street traffic frequencies, and the cadence or linguistic rhythm of peripheral chatter. Auditory contexts serve as temporal scaffolds that rhythmically bind concurrent cognitive tasks.
• Olfactory and Gustatory Context: Ambient airborne chemical signatures possess a direct, unmediated neural projection to the limbic system, bypassing the thalamic relay that processes visual and auditory stimuli. Consequently, olfactory cues are unusually potent elicitors of vivid, affectively rich autobiographical and episodic memories (a dynamic historically referred to as the Proustian phenomenon). Subtle, diffuse ambient scents, when systematically preserved between learning and testing phases, produce robust context-dependent memory effects.
• Temporal and Macro-Environmental Context: Beyond immediate sensory parameters, context extends to macro-environmental and temporal matrices. These dimensions encompass circadian timing markers (e.g., dawn versus dusk lighting characteristics), prevailing weather patterns (e.g., torrential rain versus arid sunlight filtered through windows), and seasonal atmospheric parameters. The cognitive architecture incorporates these macro-features into episodic indices, using the broader environmental era to locate and retrieve specific target traces.

2.4 Boundary Conditions and Limitations

Despite its robustness in experimental psychology, context-dependent memory is not an all-powerful, immutable law. Its influence is constrained by strict boundary conditions and cognitive parameters. One of the most consistent findings in the literature is that context-dependent effects are profoundly sensitive to the nature of the retrieval task. Environmental context effects are prominent in free recall tasks, where the individual is provided with minimal structural support and must initiate an unassisted cognitive search through memory networks. However, these effects often diminish or disappear entirely in cued recall tasks and are frequently non-existent in recognition tasks.

This dynamic is explained by the Outshining Hypothesis, articulated by Steven M. Smith. According to this framework, contextual cues are inherently weak, peripheral associative prompts. When an individual is faced with a recognition test, the presentation of the target item itself (e.g., seeing the target word printed directly on a test paper) introduces an immediate, highly potent, item-specific retrieval cue. This powerful focal cue completely “outshines” the subtle, diffuse contributions of ambient room lighting, spatial geometry, or background acoustics, rendering the external context redundant for task completion.

Furthermore, the degree of semantic processing and cognitive elaboration applied during initial learning acts as a significant buffer against environmental dependency. When an individual engages in deep, elaborative rehearsal—integrating target concepts with pre-existing semantic schemas, creating novel conceptual connections, and generating personal analogies—the resultant memory trace becomes anchored within dense, robust associative networks. This structural resilience allows the memory to be accessed through internal conceptual pathways, mitigating the need for superficial physical cues. Environmental context dependency is thus inversely correlated with the depth of semantic encoding.

Finally, the cognitive phenomenon of mental context reinstatement reveals that physical return to the original learning setting is not an absolute requirement for overcoming contextual retrieval deficits. As demonstrated by cognitive psychologists, if an individual who is tested in a novel, mismatched environment simply takes several minutes prior to the examination to vividly imagine and mentally reconstruct the physical features, sensory textures, and spatial layout of the original study room, this internal cognitive simulation can reinstate the associative pathways of the original context. This mental visualization reliably erases the statistical performance gap typically observed between matching and non-matching physical environments.

3. Comprehensive Overview of State-Dependent Memory

3.1 Theoretical Framework and Internal Milieu

State-dependent memory is fundamentally distinct from context-dependent memory in that it is governed entirely by the internal biological and psychological state of the organism. Rather than responding to the exogenous surroundings, this phenomenon relies upon the endogenous milieu: the dynamic equilibrium of neurochemical pathways, hormonal concentrations, autonomic nervous system tone, and affective states that define subjective somatic reality. State-dependent learning dictates that information acquired when the brain is functioning under a specific physiological or neuropharmacological configuration becomes linked to the unique interoceptive cues generated by that state.

The conceptual framework of state-dependent learning rests upon interoceptive conditioning and somatic integration. Just as an external environment supplies visual and auditory landmarks that are integrated into an episodic trace, the body’s internal systems continuously broadcast physiological feedback to the central nervous system. These signals encompass heart rate variability, peripheral vascular resistance, visceral muscular tension, blood glucose levels, systemic inflammation markers, and circulating stress hormones such as epinephrine and cortisol. Under the somatic marker hypothesis, these biological metrics are not mere metabolic byproducts; they form an active, dynamic internal map within the insular cortex, anterior cingulate, and amygdala that continuously contextualizes mental operations.

When information is encoded, this internal somatic state is bound to the target episodic representation. The brain’s homeostatic monitoring systems effectively utilize these somatic parameters as an indexing matrix. During retrieval, if the organism’s internal milieu has shifted significantly—such as moving from a state of acute adrenergic arousal to a state of profound physiological parasympathetic relaxation—the interoceptive markers that initially indexed the memory trace are missing. The neural pathways that require this internal hormonal and neurochemical backdrop fail to reach activation threshold, leading to retrieval failure. Consequently, the retrieval of state-dependent memories requires recreating the internal physiological environment that was present when the learning occurred.

3.2 Pharmacological and Substance-Induced Paradigms

The scientific study of state-dependent memory originated primarily in psychopharmacology rather than traditional cognitive psychology laboratories. In the mid-20th century, Donald A. Overton conducted foundational drug-discrimination experiments on non-human animal models that definitively proved the existence of drug-induced state-dependent learning. Overton demonstrated that rodents trained to escape a T-maze while under the influence of central nervous system depressants, such as phenobarbital, were completely unable to execute the learned spatial navigation choices when tested in a sober, unmedicated state. Crucially, when these animals were subsequently re-injected with the barbiturate, their functional maze navigation performance was immediately restored, demonstrating that the learned motor sequences had not decayed, but were selectively accessible only within that specific neurochemical state.

Subsequent human research quickly expanded to examine ethanol-induced state-dependent memory. In classic human alcohol paradigms, participants were instructed to memorize lists of words, complex associative pairs, or spatial arrangements while either sober or under the influence of measured doses of ethyl alcohol (typically titrated to blood alcohol concentrations around 0.08% to 0.10%). Across numerous studies, participants who encoded information while intoxicated showed poor recall when tested sober. However, when tested under identical levels of alcohol intoxication, their recall scores improved significantly compared to their sober retrieval attempts, establishing a clear state-dependent pattern.

Similar state-dependent dynamics have been documented across a wide variety of psychoactive substances, including:

• Central Nervous System Stimulants: Controlled administrations of caffeine, amphetamines, and nicotine have demonstrated state-dependent retention profiles. The heightened noradrenergic and dopaminergic signaling characteristic of stimulant states provides a distinct internal cue. Information acquired during peak stimulant activity is retrieved more efficiently if the stimulant is re-administered prior to testing.
• Sedatives, Anxiolytics, and Benzodiazepines: Clinical pharmacology has highlighted the pronounced state-dependent properties of benzodiazepines (such as diazepam, alprazolam, and triazolam). Because these agents act as positive allosteric modulators of GABA-A receptors, they significantly alter cortical excitability, frequently inducing partial anterograde amnesia while simultaneously establishing a state-dependent retrieval gate for material encoded during the drug’s therapeutic window.
• Cannabinoids: Studies assessing Delta-9-tetrahydrocannabinol (THC) have consistently documented state-dependent recall failure. The profound alteration of CB1 receptor signaling across the hippocampus and basal ganglia creates an altered neurochemical state that restricts the retrieval of information encoded under its influence when individuals return to an unmedicated baseline.

3.3 Affective States, Mood Congruence, and Arousal

State dependency is not restricted to pharmacologically induced states; it applies equally to endogenous emotional conditions and physiological arousal levels. However, in the realm of affect, cognitive science draws a critical theoretical distinction between state-dependent memory and mood-congruent memory. While these terms are frequently conflated, they refer to fundamentally distinct cognitive mechanics.

State-dependent memory in an affective context dictates that material learned in a specific emotional state (e.g., profound sadness or intense joy) is best retrieved when the individual is re-immersed in that exact emotional state, irrespective of the valence of the material itself. In pure state-dependent memory, an individual who learns a list of neutral, utilitarian facts while in a clinically depressed state will retrieve those neutral facts more effectively when depressed than when euthymic. Conversely, mood-congruent memory refers to an alignment between the emotional valence of the target material itself and the individual’s current affective state. A depressed individual preferentially recalls negative, failure-oriented, or sad life events while ignoring positive, uplifting occurrences, simply because the valence of the material matches their current mood. Gordon H. Bower integrated these dynamics into his influential Associative Network Model of Affect.

Bower posited that distinct emotions—such as joy, depression, fear, and anger—are represented within human memory architectures as central, high-order emotional nodes embedded within a broad semantic network. These affective nodes are directly wired to autonomic visceral reactions, subjective expressive behaviors, and related semantic concepts. When an emotional node is activated by real-world events or laboratory mood-induction procedures, activation spreads outwards across all connected conceptual nodes. Memories encoded during an episode of intense fear become associatively bound to that specific emotional node, meaning subsequent activation of the fear node via elevated heart rate and sympathetic tone reactivates the related episodic nodes, facilitating their retrieval.

This dynamic is illustrated in psychiatric research involving bipolar spectrum disorders. Clinical studies assessing patients cycling between manic and depressive phases have revealed pronounced state-dependent recall patterns. Experiences, interactions, and financial decisions executed during a severe manic episode are frequently inaccessible or poorly recalled during a subsequent depressive phase. However, when the patient experiences another manic episode months or years later, the episodic memories formed during the previous manic state often resurface with striking clarity and detail, driven by the recurrence of that specific neurobiological and affective state.

3.4 Neurobiological Mechanisms

The biological architecture supporting state-dependent memory relies on complex, bidirectional communication between the limbic system, interoceptive cortical regions, and neuromodulatory networks. At the center of this system is the hippocampus, the primary anatomical hub responsible for the episodic binding of divergent cortical inputs. The hippocampus is rich in receptors for classical neurotransmitters, neuropeptides, and steroid hormones. When an organism learns information, the hippocampus binds the target sensory representations processed in associative cortices to the prevailing neurochemical background.

Crucial to this binding is the amygdala, particularly the basolateral complex. The amygdala acts as an emotional and physiological gain-control mechanism. Under conditions of emotional or physiological arousal, the locus coeruleus projects directly to the basolateral amygdala, releasing high concentrations of norepinephrine. This noradrenergic surge activates intracellular cyclic adenosine monophosphate (cAMP) and protein kinase A (PKA) cascades, which in turn modulate synaptic plasticity and long-term potentiation (LTP) within the hippocampus. The strength and unique neurochemical configuration of this amygdalar activation acts as a distinct biological signature that tags the memory trace during consolidation.

Simultaneously, the insular cortex provides the continuous physiological substrate for state-dependent cues. The posterior and mid-insula continuously receive visceral sensory signals from the solitary tract and thalamus, processing blood pressure, vascular distension, gut motility, and respiratory cadence. The anterior insular cortex integrates these homeostatic signals into a conscious, subjective feeling state. When the insula’s interoceptive representation changes between encoding and retrieval, the associative pattern completion process initiated in hippocampal CA3 networks encounters a signal mismatch, impeding the reactivation of the target trace.

Finally, glucocorticoid receptor dynamics within the prefrontal cortex (PFC) exert strong control over retrieval success. During stress, the adrenal cortex secretes high volumes of cortisol, which crosses the blood-brain barrier and binds to two main receptor types: mineralocorticoid receptors (MRs) and glucocorticoid receptors (GRs). When cortisol levels match between encoding and retrieval, the baseline firing rates and excitability of prefrontal pyramidal neurons remain structurally aligned. However, if an individual encounters acute stress during retrieval that was not present during learning, the sudden saturation of GRs in the prefrontal cortex can disrupt executive retrieval processing, preventing the PFC from properly coordinating the memory search.

4. Key Differences Between Context-Dependent and State-Dependent Memory

4.1 Locus and Nature of Cues: External Environment vs. Internal Milieu

The primary distinction between context-dependent and state-dependent memory lies in the spatial locus and fundamental nature of the informational cues governing retrieval. Context-dependent memory relies entirely upon exogenous cues—informational inputs that originate outside the organism’s physical boundary. These cues are public, objective, and shared. A classroom, a research laboratory, or an outdoor environment can be simultaneously experienced, measured, and verified by multiple independent observers. The sensory pathways that capture these contextual cues are exteroceptive: photons hitting the retina, sound waves vibrating the tympanic membrane, and ambient molecules interacting with olfactory receptors.

Conversely, state-dependent memory is governed entirely by endogenous cues—signals that originate within the organism’s internal somatic boundary. These cues are fundamentally private, subjective, and interoceptive. No outside observer can directly experience another individual’s elevated sympathetic arousal, drug-induced neurochemical shift, or specific somatic fatigue. State-dependent retrieval relies on interoceptive and somatosensory pathways: baroreceptors, chemoreceptors, visceral vagal afferents, and direct neurochemical interactions at central receptor sites. Measuring context requires physical instrumentation such as lux meters, sound level meters, spectrophotometers, and spatial coordinate mapping. Measuring state requires biological and physiological instrumentation: blood serum assays, electroencephalography (EEG), heart rate variability analysis, galvanic skin conductance, and endocrine monitoring.

4.2 Volitional Control and Environmental Manipulation

The operational difference between these two phenomena is further underscored by the varying degrees of volitional control an individual can exercise over external contexts versus internal states. External physical contexts can generally be manipulated, restructured, or exited through intentional motor action. If a student or professional realizes that their physical surroundings do not support the recall of previously learned material, they can physically relocate to an environment that more closely mirrors their original study location, or alter variables such as room lighting, desk positioning, or acoustic inputs.

Internal physiological states, however, are largely governed by the autonomic nervous system and neurochemical dynamics, making them resistant to rapid, precise, and volitional control. An individual cannot instantly shift their blood cortisol levels, spontaneously reset their blood alcohol concentration to zero, or immediately normalize catecholamine surges through sheer willpower. Inducing or terminating a physiological state often requires significant time, behavioral interventions, or pharmacological agents, which introduces substantial ethical, logistical, and biological overhead. In experimental paradigms, manipulating environmental context involves non-invasive, ethically benign actions (e.g., changing rooms, swapping background music, or adjusting screen brightness). In contrast, manipulating state often requires invasive procedures: administering psychoactive substances, running participants to exhaustion on treadmills, or inducing genuine psychological stress, all of which carry profound ethical constraints and potential health risks.

4.3 Asymmetry and Directionality of Effects

A major theoretical divergence between context-dependent and state-dependent memory lies in the phenomenon of directional asymmetry. In classical context-dependent memory experiments, retrieval decrements tend to be largely symmetrical across reciprocal environmental shifts. If an individual learns material in Room A and is tested in Room B, the decrement in recall performance is typically comparable to that of an individual who learns in Room B and is tested in Room A. The primary operational variable is simply the mismatch between the physical spaces, without one room universally degrading the brain’s executive processing capacity.

State-dependent memory, particularly in pharmacological and affective contexts, frequently exhibits strong, predictable directional asymmetry. Specifically, transitioning from an altered or intoxicated internal state at encoding to a normal or sober state at retrieval typically produces a much more severe retrieval deficit than the reverse transition (learning sober and testing altered). For example, if a participant learns a word list while heavily intoxicated by alcohol and is subsequently tested while completely sober, recall performance drops precipitously. However, if a participant learns while sober and is tested while intoxicated, the retrieval deficit—while still present due to cue mismatch—is often significantly less severe.

This asymmetry occurs because state shifts, unlike benign environmental shifts, regularly alter overall cognitive and executive processing capacity. Central nervous system depressants, extreme affective states, and high stress levels induce structural cognitive changes—such as alcohol myopia, restricted attentional focus, and prefrontal cortex down-regulation. When material is encoded in an altered state, the original encoding is often fragmented, shallow, and weakly consolidated. Testing in a sober state removes the few interoceptive cues that were bound to the trace, exposing its structural weaknesses. In contrast, information encoded in a sober, optimal cognitive state undergoes deep, robust semantic elaboration. While an altered state at testing may impair the individual’s ability to search memory, the underlying trace remains robust, allowing some retrieval to occur through sheer trace strength.

4.4 Task Sensitivity and Retrieval Metrics

The comparative persistence of context-dependent versus state-dependent memory across different retrieval tasks reveals fundamentally distinct cognitive architectures. Context-dependent memory effects are fragile and task-sensitive. As established by the Outshining Hypothesis, context-dependent effects typically vanish when testing shifts from free recall to cued recall or recognition tasks. Because external context cues are diffuse and peripherally bound via the parahippocampal cortex, the presentation of a direct, target-specific cue (such as the actual word in a recognition test) immediately overrides and renders irrelevant the external environmental background.

State-dependent memory effects, however, display greater resilience across varied testing tasks. While state dependency is also strongest in free recall, significant state-dependent performance deficits frequently persist even during cued recall and complex recognition paradigms. This heightened persistence stems from the fact that internal physiological states do not simply provide localized associative cues; they systematically alter global cortical excitability, neurotransmitter availability, and information processing modes. An internal state characterized by profound anxiety or severe depressant intoxication alters the functional connectivity between the prefrontal cortex and hippocampal formation. As a result, the brain’s capacity to discriminate between target stimuli and familiar lures during recognition tasks is genuinely compromised by the physiological state itself, preventing external target cues from entirely outshining the internal state mismatch.

When to Focus on Context vs. State

Understanding when to target context-dependent versus state-dependent interventions depends upon the cognitive goals and the practical constraints of the operational environment:

• Prioritize Context-Dependent Strategies When:
  • The physical environment where performance will be tested is highly predictable and structurally standardized (e.g., standard testing centers, flight simulators, or standardized operational environments).
  • Learners have the logistical freedom to deliberately vary their study environments across multiple rooms and locations, thereby preventing brittle, hyper-specific contextual binding.
  • Physical objects, specialized equipment, or spatial configurations are central to the task, making visual and spatial cues the primary anchors for memory recall.
  • Interventions must remain non-invasive, cost-effective, and completely free of pharmacological or autonomic disruption.
• Prioritize State-Dependent Strategies When:
  • Performance will occur under extreme psychological or physiological stress, such as military combat, emergency trauma surgery, or emergency aviation procedures. In these domains, matching the elevated autonomic arousal during training is essential for real-world execution.
  • Designing psychotherapeutic treatments for post-traumatic stress or anxiety disorders, where extinction learning must transfer successfully from a calm clinical office to emotionally turbulent real-world settings.
  • Evaluating the legal reliability of eyewitnesses or suspects who were heavily intoxicated, terrified, or sleep-deprived during a crime.
  • Managing substance use disorders, where internal craving states and withdrawal-induced somatic changes trigger drug-seeking recall pathways.

5. Similarities and Theoretical Overlap

5.1 Rootedness in the Encoding Specificity Principle

Despite their physical and physiological differences, context-dependent and state-dependent memory are unified by a single foundational architecture: the Encoding Specificity Principle. Both paradigms demonstrate that human memory retrieval is fundamentally non-exhaustive, reconstructive, and contingent on cue availability. Neither context-dependent nor state-dependent memory constitutes an independent, structurally isolated memory system. They do not operate through distinct, dedicated anatomical repositories comparable to the structural divide between declarative (explicit) and non-declarative (implicit) memory.

Instead, both phenomena represent different behavioral expressions of the exact same basic retrieval mechanism. Both validate the core psychological axiom that forgetting is primarily a failure of accessibility, not an immediate loss of availability. When a scuba diver fails to recall terrestrial word lists underwater, or a sober individual struggles to recall events from an alcohol-induced blackout, the memory trace itself is rarely obliterated. Instead, the information remains stored within structural neocortical ensembles, yet is temporarily inaccessible due to an absence of the appropriate cue pattern needed to trigger hippocampal pattern completion. Both phenomena confirm that episodic memory traces are holistic, configural records composed of both the focal item and the multivariant background features present during acquisition.

5.2 Cognitive Load and Attentional Binding

A second fundamental parallel is their shared vulnerability to cognitive load and attentional modulation during the encoding phase. In both paradigms, the integration of background cues—whether the color of a wall or a change in heart rate—occurs predominantly via incidental encoding. Human working memory possesses strictly finite processing capacity. Under conditions where an individual’s cognitive load is completely consumed by a demanding, highly complex primary task, the brain’s capacity to bind peripheral contextual and somatic features decreases significantly.

When working memory is overloaded, hippocampal pattern separation and relational binding mechanisms selectively prioritize focal, task-critical information at the expense of ambient cues. Consequently, when an individual engages in deep, elaborative, semantic processing—investing conscious effort into evaluating meaning, thematic structures, and logical relationships—the reliance on both external environmental cues and internal somatic cues is markedly reduced. Furthermore, both context-dependent and state-dependent cues exhibit comparable temporal decay profiles; as the retention interval between learning and testing extends over weeks or months, the specific peripheral associations that bind the memory trace to its original context or state naturally degrade, leaving only the consolidated semantic core.

5.3 Synergistic and Interactive Co-Occurrence

While experimental researchers intentionally isolate external context from internal state to satisfy scientific reductionism, ecological human experience rarely decouples them. In naturalistic settings, context-dependent and state-dependent cues operate in continuous synergy, functioning as an integrated, compound retrieval cue matrix. Specific physical contexts reliably trigger predictable physiological and affective states. A student who enters an intimidating, windowless examination room does not experience an isolated change in physical surroundings; that specific space immediately elicits elevated sympathetic nervous system activity, autonomic tachycardia, and anxiety.

When an individual experiences retrieval failure in that examination room, the deficit is driven by an unseparated combination of mismatched external cues (different from their home study space) and mismatched internal cues (a panic state contrasting with their calm study mindset). Furthermore, the cognitive intervention of mental context reinstatement often inadvertently reinstates the corresponding internal physiological state. When a person vividly visualizes their calm home office, that mental imagery can trigger a parasympathetic down-regulation of heart rate, thereby supplying both the missing external visual cues and the original internal physiological baseline needed for successful recall.

6. Real-World Applications and Applied Implications

6.1 Educational Practice and Academic Strategy

The practical implications of cue-dependent retrieval in educational contexts are profound. One of the most common pitfalls encountered by students is the “bedroom study trap.” Many learners complete the bulk of their academic revision in an overly relaxed, familiar physical environment: lounging in bed or at a casual desk, wearing comfortable clothes, sipping tea, listening to familiar background music, and maintaining a low-stress physiological baseline. Under these conditions, the learned academic material becomes incidentally bound to those specific external surroundings and that calm internal state.

When these students subsequently enter a formal examination hall, they experience a double retrieval cue mismatch. The physical environment shifts to an expansive, brightly lit, silent room filled with hundreds of peers, while their internal physiological state shifts from low-stress relaxation to high-stress sympathetic arousal characterized by elevated cortisol and heart rate. This abrupt transformation can trigger acute retrieval failure, commonly referred to as “choking” or blanking out on an exam. The information is not gone, but the cues required to access it have been removed.

To overcome this vulnerability, cognitive science advocates two primary interventions:

• Contextual and State Variation: Rather than studying in a single, hyper-consistent setting, students should intentionally vary their learning environments across multiple locations (e.g., quiet library floors, active coffee shops, shared classrooms, and home offices) and varied physiological states (e.g., varying times of day, different caffeine levels). This variable encoding strategy breaks the brittle, incidental binding between the academic material and any single environment or internal state, forcing the brain to construct a robust, generalized semantic representation that can be accessed anywhere.
• Deliberate State-Matching: If contextual variation is not feasible, students should practice intentional state-matching. This involves aligning caffeine consumption, sleep architecture, and testing times between revision sessions and exam day. Furthermore, students can be trained in mental context reinstatement: spending sixty seconds at the start of an examination closing their eyes and vividly reconstructing their primary study space, thereby mentally reactivating the necessary episodic pathways.

6.2 Clinical Psychology, Psychotherapy, and Trauma

In clinical psychology and psychopathology, the principles of state-dependent and context-dependent memory are essential for understanding trauma and optimizing therapeutic interventions. Post-Traumatic Stress Disorder (PTSD) provides a clear example of state-dependent intrusion. When an individual survives a traumatic event, the terrifying sensory experience is consolidated under conditions of intense noradrenergic and glucocorticoid surges, which strongly activate the basolateral amygdala. The resulting traumatic memory trace is deeply bound to that extreme state of autonomic hyperarousal.

Months or years later, if the individual encounters an unrelated everyday stressor that causes a sudden spike in sympathetic tone—such as intense physical exertion, an unexpected medical shock, or acute panic—the internal surge in adrenaline can act as an interoceptive retrieval cue. This somatic shift can trigger involuntary state-dependent retrieval of the traumatic memory, manifesting as a vivid, intrusive flashback. The patient’s cognitive system, detecting the return of the traumatic physiological signature, completes the pattern by surfacing the original episodic trauma.

Conversely, state- and context-dependency can undermine standard psychotherapy through the phenomenon of extinction renewal. In prolonged exposure therapy for anxiety and phobias, a patient repeatedly confronts their feared stimulus in a safe, predictable, and supportive environment: the therapist’s office. Over multiple sessions, the patient successfully acquires new inhibitory learning (extinction), and their fear response drops to zero. However, when the patient encounters that same phobic stimulus out in the real world (a novel physical context) while experiencing unexpected fatigue or stress (an altered internal state), the extinguished fear response often returns with full force. The newly acquired safety memory was context-dependent, locked to the therapist’s calm office, allowing the original fearful association to re-emerge elsewhere.

To ensure long-term clinical efficacy, modern cognitive-behavioral therapists intentionally vary the physical contexts of exposure therapy, conducting sessions in real-world environments, parks, and busy public transit centers. Furthermore, therapists use state-variation techniques, having patients deliberately elevate their heart rates through physical exercise or hyperventilation during therapy. By anchoring coping strategies across diverse external environments and elevated physiological states, clinicians ensure that therapeutic gains survive beyond the clinic door.

6.3 Forensic Investigations and Eyewitness Testimony

The forensic domain and criminal justice systems are continuously shaped by the limitations and mechanics of cue-dependent retrieval. Eyewitness testimony is notoriously susceptible to retrieval failure and confabulation when formal law enforcement questioning occurs in sterile, standard police interrogation rooms. The stark physical contrast between the crime scene and the interrogation facility removes the ambient external cues that were present during the criminal event, limiting the witness’s capacity to access critical episodic details.

To overcome this limitation, forensic psychologists Ronald P. Geiselman and Edward P. Fisher developed the Cognitive Interview technique, an evidence-based questioning methodology designed explicitly to leverage mental context reinstatement without contaminating witness memory. Rather than asking rapid, direct questions, the interviewer guides the witness to close their eyes and mentally reconstruct the physical environment and subjective state present at the time of the crime. The witness is prompted to visualize the environmental lighting, the weather, the spatial placement of objects, the background sounds, and their own internal emotional feelings, thoughts, and physiological reactions during the event. By systematically reinstating both the external context and the internal state in the mind’s eye, the Cognitive Interview routinely yields significantly more accurate, non-confabulated factual details than traditional police interrogation approaches.

Furthermore, forensic science must navigate the legal dilemmas surrounding witness and suspect substance intoxication. A witness who observes a violent altercation while heavily intoxicated by alcohol may demonstrate poor recall when questioned sober by detectives the following morning. Defense and prosecution attorneys frequently debate the reliability of such testimony. Understanding the directional asymmetry of state-dependent memory reveals that while intoxication at encoding impairs trace depth, returning the individual to a slightly elevated physiological state or leveraging intense mental reconstruction can sometimes recover details that appear permanently lost during sober questioning.

6.4 High-Stakes Occupational and Military Training

In high-stakes occupations—such as military special operations, commercial aviation, aerospace engineering, nuclear powerplant management, and emergency trauma surgery—the failure of retrieval under crisis conditions can be catastrophic. When an emergency strikes, operators are exposed to extreme auditory alarms, chaotic visual fields, and sudden surges in autonomic arousal, including tachycardia exceeding 160 beats per minute, peripheral vasoconstriction, and cortisol spikes.

If these operators have been trained exclusively through calm, didactic lectures or low-stress desktop computer simulations, their knowledge remains bound to a relaxed internal and external baseline. When thrust into an actual life-or-death crisis, they frequently experience state-dependent retrieval collapse. To counteract this, modern elite training regimens employ two foundational strategies:

• Stress Inoculation Training (SIT): Military, law enforcement, and emergency medical personnel undergo rigorous simulations designed to intentionally induce severe physiological arousal. By introducing intense acoustic noise, smoke, physical exhaustion, time pressure, and simulated danger, instructors force trainees to practice decision-making, procedural recall, and motor skills while under extreme autonomic stress. This binds survival-critical procedures directly to the physiological state of high arousal, ensuring that when a real emergency triggers an adrenaline surge, the state serves as a positive retrieval cue rather than a disruptive barrier.
• High-Fidelity Contextual Simulation: Aviation and aerospace domains rely on full-motion, high-fidelity flight simulators that replicate cockpit geometry, switch resistance, visual horizon graphics, and audio communication channels down to the smallest detail. By maximizing the physical contextual overlap between the training environment and the real-world operational space, organizations minimize transfer deficits and ensure that pilot responses remain rapid, accurate, and automatic.

7. Frequently Asked Questions

7.1 Can Context-Dependent and State-Dependent Cues Operate Simultaneously?

Yes. In natural, real-world conditions, context-dependent and state-dependent cues almost always operate simultaneously as an integrated, compound retrieval cue matrix. When an individual learns and recalls information in the real world, the external physical environment (sights, sounds, room layout) and the internal somatic state (hormones, heart rate, emotional affect) are encoded together into the episodic trace. When both external surroundings and internal states match the encoding conditions, retrieval probability reaches its empirical maximum.

If these cues conflict—for example, if an individual is tested in the exact same physical room where they studied, but is experiencing an acute panic attack that was entirely absent during learning—the cognitive system encounters mixed retrieval signals. In such cases, memory performance typically falls to an intermediate level, demonstrating that the presence of matching external physical cues can partially, but not entirely, compensate for the retrieval deficits caused by mismatched internal physiological states.

7.2 How Does Mood-Congruent Memory Differ from State-Dependent Memory?

Although both concepts involve emotion, they describe fundamentally different cognitive mechanisms. State-dependent memory depends entirely on the match between the individual’s internal emotional state at the time of encoding and their state at retrieval, regardless of the emotional valence of the material being learned. For example, if you memorize a list of neutral, non-emotional mathematical formulas while feeling intensely sad, state-dependent memory predicts you will recall those neutral formulas best when you are sad again.

Mood-congruent memory, on the other hand, describes an alignment between the emotional valence of the target information itself and the individual’s current mood state at retrieval. In this case, an individual who is currently feeling sad or depressed will preferentially recall negative, depressing life events (such as a past breakup or personal failure) rather than happy memories, simply because the negative tone of the information matches their current emotional state. State dependency focuses on the congruence of physiological and neurochemical states across time; mood congruence focuses on the congruence between an individual’s current mood and the emotional tone of the material itself.

7.3 Is Context-Dependent Memory Stronger for Recall than for Recognition?

Yes, empirical research consistently demonstrates that context-dependent memory effects are substantially stronger in free recall tasks and often disappear completely in recognition tasks. This variance is explained by the Outshining Hypothesis. In a free recall task, an individual receives minimal external guidance and must initiate a self-directed search through memory networks. Under these conditions, subtle, peripheral cues—such as room geometry, ambient sounds, or scents—can provide the necessary activation needed to locate and retrieve the stored memory trace.

In contrast, a recognition task provides the individual with the target item directly (e.g., asking “Was the word ‘ELEPHANT’ on the list?”). The presentation of the actual target item supplies an immediate, highly salient, item-specific retrieval cue that directly activates the relevant semantic and episodic node. This powerful focal cue completely “outshines” the weak, diffuse contextual features of the surrounding room, rendering the ambient physical environment functionally irrelevant for task completion.

7.4 Can You Artificially Recreate Internal States to Aid Retrieval?

Yes, within certain physiological and practical boundaries, internal states can be deliberately recreated to enhance retrieval. Controlled physiological adjustments—such as consuming an equivalent dose of caffeine to match study arousal, performing brief physical exercise to match a faster heart rate, or utilizing paced breathing techniques to replicate a state of parasympathetic relaxation—can successfully reinstate key interoceptive cues, aiding memory recall.

Similarly, psychological mood-induction techniques, such as listening to specific musical selections or engaging in targeted emotional reflection, can reliably recreate the affective state present during initial learning. While pharmacologically reinstating drug states (such as re-ingesting sedatives or alcohol) is theoretically possible under state-dependent models, it is practically counterproductive for everyday academic or professional tasks, as the substance’s direct cognitive side effects generally impair executive function far more than the state match helps it. Modern somatic tracking devices and biofeedback tools provide practical, non-invasive methods for monitoring and adjusting autonomic arousal to optimize retrieval.

8. Conclusion and Synthesis

8.1 Summary of Distinctions and Synergies

The distinction between context-dependent memory and state-dependent memory reflects the fundamental division between the world outside the skin and the world within it. Context-dependent memory highlights the influence of the exogenous environment, showing how ambient visual landmarks, acoustic profiles, spatial dimensions, and olfactory signatures are bound to episodic memory traces. As demonstrated by classic experiments—from Godden and Baddeley’s underwater divers to modern virtual reality paradigms—environmental congruence supports retrieval, particularly when cognitive search is unassisted. However, these external cues are largely symmetrical in their effects, accessible to outside observers, and susceptible to being outshone by focal, item-specific cues during recognition tasks.

State-dependent memory, by contrast, centers upon the endogenous biological milieu: the neurochemical, hormonal, autonomic, and affective states that define internal somatic reality. Rooted in psychopharmacology, affective neuroscience, and interoceptive monitoring, state dependency reveals that the brain utilizes internal physiological configurations as indexing matrices for stored representations. Unlike external context, state shifts frequently exhibit marked directional asymmetry, persist across more robust testing conditions due to systemic changes in cortical excitability, and remain private, subjective, and resistant to rapid, volitional control.

Yet, despite these differences, both phenomena share a theoretical foundation under Tulving’s Encoding Specificity Principle. Both demonstrate that forgetting is often a failure of accessibility rather than total trace decay, and both prove that human memory traces are holistic, configural records composed of the target item and the background conditions present during learning. In everyday life, external contexts and internal states continuously interact. Environments elicit predictable emotional and physiological responses, generating compound cue matrices that govern our ability to access stored knowledge.

8.2 Future Directions in Cognitive Retrieval Research

As cognitive psychology and neuroscience advance, the investigation of context-dependent and state-dependent retrieval continues to evolve through new technologies and methodologies. Advanced functional neuroimaging (high-resolution fMRI and magnetoencephalography) combined with optogenetic techniques in animal models is beginning to map the precise neural circuits that link the insular cortex, basolateral amygdala, and entorhinal-hippocampal networks. These advances allow researchers to observe the pattern completion process in real time, showing how the brain integrates interoceptive somatic markers and exteroceptive spatial maps during memory consolidation and retrieval.

Concurrently, the rapid rise of immersive digital technology—including high-fidelity virtual reality (VR), augmented reality (AR), and spatial computing—is redefining the boundaries of environmental context. Researchers can now manipulate virtual environments with absolute precision, systematically testing which visual, spatial, and acoustic features are necessary to trigger context-dependent retrieval. Furthermore, neurotechnological interventions targeting sleep-dependent memory consolidation—such as Targeted Memory Reactivation (TMR)—are exploring how subtle sensory cues (such as ambient scents or auditory tones) presented during slow-wave sleep can selectively reactivate and strengthen specific memories, bridging external context with internal brain states.

Finally, the integration of wearable biometric sensors, real-time physiological tracking, and adaptive machine learning models is opening the door to personalized learning systems. In the coming decades, educational software and occupational training platforms may dynamically assess a learner’s pupil dilation, heart rate variability, galvanic skin response, and electroencephalographic signatures in real time. By adjusting digital environments and instructional pacing to match the optimal physiological states required for long-term retention, future technologies may harness the principles of context- and state-dependent memory, transforming our understanding of human learning and cognitive potential.

References

• Anderson, J. R. (1983). The architecture of cognition. Harvard University Press.
• Baddeley, A. D. (1997). Human memory: Theory and practice (Revised ed.). Psychology Press.
• Bower, G. H. (1981). Mood and memory. American Psychologist, 36(2), 129–148. https://doi.org/10.1037/0003-066X.36.2.129
• Damasio, A. R. (1994). Descartes’ error: Emotion, reason, and the human brain. G.P. Putnam’s Sons.
• Eich, J. E. (1980). The cue-dependent nature of state-dependent retrieval. Memory & Cognition, 8(2), 157–173. https://doi.org/10.3758/BF03213419
• Fisher, R. P., & Geiselman, R. E. (1992). Memory-enhancing techniques for investigative interviewing: The cognitive interview. Charles C Thomas, Publisher.
• Geiselman, R. E., Fisher, R. P., MacKinnon, D. P., & Holland, H. L. (1985). Eyewitness memory enhancement in the police interview: Cognitive retrieval mnemonics versus hypnosis. Journal of Applied Psychology, 70(2), 401–412. https://doi.org/10.1037/0021-9010.70.2.401
• Godden, D. R., & Baddeley, A. D. (1975). Context-dependent memory in two natural environments: On land and underwater. British Journal of Psychology, 66(3), 325–331. https://doi.org/10.1111/j.2044-8295.1975.tb01468.x
• Overton, D. A. (1964). State-dependent or “dissociated” learning produced with pentobarbital. Journal of Comparative and Physiological Psychology, 57(1), 3–12. https://doi.org/10.1037/h0040512
• Rasch, B., Büchel, C., Gais, S., & Born, J. (2007). Odor cues during slow-wave sleep prompt declarative memory consolidation. Science, 315(5817), 1426–1429. https://doi.org/10.1126/science.1138581
• Smith, S. M. (1988). Environmental context-dependent memory. In G. M. Davies & D. M. Thomson (Eds.), Context and memory (pp. 13–34). John Wiley & Sons.
• Smith, S. M., Glenberg, A., & Bjork, R. A. (1978). Environmental context and human memory. Memory & Cognition, 6(4), 342–353. https://doi.org/10.3758/BF03197465
• Tulving, E., & Thomson, D. M. (1973). Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80(5), 352–373. https://doi.org/10.1016/0010-0285(73)90017-8