Atkinson-Shiffrin Model of Memory (1968)
The Atkinson-Shiffrin model, proposed in 1968, is one of the earliest and most influential models of memory. It describes memory as a linear process involving three separate stores: |
🔁 Sensory Register → Short-Term Memory → Long-Term Memory |
It emphasizes encoding, storage, and retrieval as the core processes of memory. |
The Three Memory Stores |
Sensory Register
<1–2 seconds
Very large Raw/unprocessed (modality-specific: visual, auditory, etc.)
Rapid decay |
Short-Term Memory (STM)
~15–30 seconds
7 ± 2 items (Miller, 1956) Acoustic (mainly)
Displacement & decay |
Long-Term Memory (LTM)
Potentially lifetime
Unlimited
Primarily semantic
Retrieval failure, interference |
Key Processes in the Model |
Attention
Focusing on specific sensory input
Moves info from sensory to STM |
Rehearsal
Repeating information mentally or aloud
Transfers info from STM to LTM |
Encoding
Transforming input for storage
STM: acoustic; LTM: semantic |
Retrieval
Accessing stored information
From LTM back to STM for use |
Forgetting
Loss of stored info
Each store has different causes (e.g., decay, interference) |
Strengths of the Model |
✅ Clear structure—easy to test experimentally |
✅ First to distinguish memory types systematically |
✅ Explains serial position effect |
✅ Supported by neuropsychological evidence (e.g., patient HM) |
Criticisms & Limitations |
❌ Oversimplified – memory is not purely linear |
❌ Too focused on rehearsal – not the only route to LTM |
❌ Doesn’t explain implicit memory or procedural learning |
❌ Lacks explanation of interaction between STM and LTM (e.g., chunking uses LTM knowledge in STM) |
Baddeley & Hitch’s Working Memory Model (1974)
1. Why It Was Proposed |
To replace the oversimplified Short-Term Memory (STM) store in Atkinson & Shiffrin's model.
Emphasized that memory is not a single passive store, but an active, multi-component system for holding and manipulating information. |
Core Components of the Model |
a. Central Executive |
💡 Main control system
Directs attention, allocates tasks to subsystems.
Has limited capacity, doesn’t store info itself.
Involved in planning, problem-solving, decision-making. |
b. Phonological Loop |
Deals with verbal/auditory information.
Two sub-parts:
Phonological Store ("inner ear") – holds spoken words briefly.
Articulatory Control Process ("inner voice") – allows rehearsal.
Crucial for language processing and learning. |
c. Visuo-Spatial Sketchpad |
Handles visual and spatial information.
Called the "inner eye".
Involved in navigation, mental imagery, and visual memory.
Later split into:
Visual cache (stores form/color)
Inner scribe (records spatial/movement info) |
d. Episodic Buffer (added in 2000) |
Integrates info from PL, VSS, and LTM into coherent episodes.
Has limited capacity.
Useful in working with integrated multi-modal information (e.g., stories). |
Supporting Research & Evidence |
🧠 Dual-Task Studies (Baddeley & Hitch, 1974) Participants performed two tasks at once:
One verbal (e.g., repeating numbers)
One reasoning (e.g., true/false questions)
Result: Could do both, but slower → suggests separate systems (not a single STM).
🧪 Word Length Effect (Baddeley et al., 1975) Short words are recalled better than long words.
Supports idea of a time-limited phonological loop.
🎨 Logie (1995) Gave evidence for separate visual and spatial stores in the visuo-spatial sketchpad.
🧍♂️ KF Case Study (Shallice & Warrington, 1970) Brain damage: poor verbal STM, good visual memory.
Supports the existence of different STM components. |
Strengths of the Model |
Explains multi-tasking.
Evidence from brain imaging (e.g., different areas for verbal/visual tasks).
More realistic than the MSM – reflects cognitive flexibility.
Accounts for active processing (not just storage). |
Weaknesses of the Model |
Central Executive is vague – lacks detailed explanation.
Little is known about how subsystems interact.
Mostly tested in lab settings – ecological validity?
May underestimate the role of LTM in working memory tasks. |
Craik and Lockhart’s Levels of Processing Model
Overview and Key Concepts |
Craik and Lockhart challenged the multi-store model of memory.
Proposed that memory is a by-product of the depth of processing, not of distinct stores.
Emphasis is on how information is processed, not where it is stored.
Deeper processing = better long-term retention.
Memory durability depends on levels of analysis (not repetition alone).
|
🔍 Levels of Processing |
Shallow Processing
Focuses on surface features (e.g., structure, sound).
Includes: visual (what it looks like) and phonemic (how it sounds) encoding.
Results in weak, short-lived memory traces.
Intermediate Processing
Involves some analysis, such as recognizing a word's sound or rhyme.
Better than shallow, but still not optimal for long-term retention.
Deep (Semantic) Processing
Focuses on meaning, context, or relating new info to existing knowledge.
Encourages elaboration, association, and comprehension.
Produces stronger, more durable memory traces.
|
🧪 Supporting Experiments
|
Craik & Tulving (1975)
Participants were asked questions about words requiring different depths of processing:
Shallow (Is the word in capital letters?)
Intermediate (Does it rhyme with ‘cat’?)
Deep (Does it fit in the sentence: “He met a ___ on the street”?)
Findings: Words processed deeply were recalled more accurately.
Conclusion: Depth of processing has a direct effect on memory.
|
✔️ Strengths of the Model
|
Explains why elaborative rehearsal is more effective than maintenance rehearsal.
Emphasizes cognitive processes over storage structures.
Supported by a range of experimental evidence.
Influential in educational practices – encouraged meaningful learning.
|
❌ Limitations of the Model
|
No clear definition of what counts as “depth” – it's vague and circular.
Difficult to objectively measure levels of processing.
May underestimate the role of memory structures (e.g., STM vs. LTM distinction).
Doesn’t explain why deep processing doesn’t always lead to better recall.
|
🧠 Applications
|
Learning techniques: Encourages elaboration, summarization, and connecting to prior knowledge.
Useful in designing educational content for better retention.
Applied in understanding encoding processes in memory disorders.
|
Tulving’s LTM Model
📍1. Introduction
Proposed by Endel Tulving in 1972 and revised in 1985.
Argued that LTM is not a single store, but consists of distinct subsystems.
First to clearly separate Episodic and Semantic memory; later added Procedural and Priming.
|
🧠 2. Main Components of Long-Term Memory
a. Episodic Memory
Stores personal experiences tied to a specific time and place.
Example: Remembering your last birthday.
Context-dependent and involves mental time travel.
Neural basis: Hippocampus, medial temporal lobe.
b. Semantic Memory
Stores general knowledge, facts, concepts, and meanings.
Example: Knowing that Paris is the capital of France.
Not linked to personal experience or time.
Neural basis: Temporal lobe, especially left hemisphere structures.
c. Procedural Memory (added later)
Memory for skills and actions; often unconscious.
Example: Riding a bicycle, typing on a keyboard.
Neural basis: Cerebellum, motor cortex, basal ganglia.
d. Priming (also called Perceptual Representation System)
Implicit memory where exposure to one stimulus influences response to another.
Example: More likely to recognize a word you've seen recently.
Neural basis: Neocortex, visual association areas.
|
🧪 3. Supporting Evidence
KC (Tulving, 1989): Brain injury left him with no episodic memory but intact semantic memory.
Clive Wearing: Severe amnesia; lost episodic memory but retained procedural skills (e.g., piano playing).
Neuroimaging: PET and fMRI scans show different brain regions activate for episodic vs. semantic tasks.
|
✔️ 4. Strengths of the Model
Explains different types of LTM observed in brain-damaged patients.
Supported by neuropsychological and brain imaging evidence.
Provides a more realistic, detailed view of memory compared to older models.
Accounts for both conscious (explicit) and unconscious (implicit) memory.
|
❌ 5. Limitations of the Model
Overlap between types of LTM (e.g., semantic memories often have episodic origins).
Difficult to clearly separate memory systems experimentally.
Not all memories fit neatly into just one category.
|
🧠 6. Applications
Understanding amnesia, Alzheimer’s, and other memory disorders.
Applied in education, as episodic memory can help encode semantic content.
Used in therapeutic approaches for trauma and skill training.
|
Parallel Distribution Processing Model
Introduction and Overview
|
Developed in the 1980s by researchers like Rumelhart, McClelland, and the PDP Group.
Also known as Neural Network Model or PDP (Parallel Distributed Processing) Model.
Inspired by how neurons function in the brain.
Emphasizes distributed, parallel processing of information across a network.
|
Key Concepts
|
Units: Basic processing elements that simulate neurons.
Connections: Like synapses between neurons; can be strong, weak, excitatory, or inhibitory.
Nodes: Represent concepts, features, or word meanings.
Activation: When a node or unit is “turned on” by incoming information.
Spreading Activation: When activation spreads across the network to related nodes.
Weighting: Each connection has a “weight” which affects how signals are processed.
Learning: Occurs through adjustment of connection weights (Hebbian learning principles: “cells that fire together, wire together”).
|
🧠 How Memory Works in This Model
|
Memory is not stored in one place, but is distributed across a network.
Each memory is represented by a pattern of activation across multiple nodes.
Retrieval is reconstructive – patterns of activation are recreated rather than replayed exactly.
More overlapping patterns = more associations = easier retrieval.
Forgetting occurs when activation patterns become weak or disrupted.
|
🧪 Supporting Evidence and Applications
|
Simulations show the model can learn language, recognize patterns, and even generalize to new inputs.
Explains phenomena like tip-of-the-tongue, semantic priming, and graceful degradation (partial memory loss).
Has influenced fields like AI, cognitive neuroscience, and psycholinguistics.
|
✔️ Strengths of the Model
|
Biologically inspired – mirrors how the brain likely processes information.
Explains how learning and memory are adaptive and flexible.
Can account for partial recall, generalization, and error patterns in memory.
Describes how we process meaning, not just store information.
|
❌ Limitations of the Model
|
Often too abstract or complex to fully map onto actual brain activity.
Difficult to test and falsify experimentally.
Sometimes fails to distinguish between different memory types (e.g., episodic vs. semantic).
May oversimplify cognitive functions by focusing only on activation patterns.
|
Encoding in Memory
📌 What is Encoding? |
Encoding refers to the initial process of transforming sensory input into a form that can be stored in the brain.
It is the first stage of the memory process (Encoding → Storage → Retrieval).
Encoding determines the strength, durability, and accessibility of memory traces.
It is not passive—how we encode influences how well we remember.
|
🧠 Types of Encoding |
Visual Encoding: Based on the appearance of stimuli (e.g., images, shapes, colors).
Acoustic Encoding: Based on the sound of information (e.g., rhymes, rhythm, verbal repetition).
Semantic Encoding: Based on meaning; involves elaboration and association with existing knowledge.
Tactile Encoding: Based on physical sensations (e.g., texture).
Olfactory and Gustatory Encoding: Rare, but potent when linked with emotional or episodic memories.
|
🔍 Levels of Processing Theory (Craik & Lockhart, 1972) |
Memory is influenced more by depth of processing than by separate memory stores.
Shallow processing: Structural and phonemic processing leads to weak memory traces.
Deep processing: Semantic encoding leads to stronger and more durable memory.
Depth is enhanced by elaboration, distinctiveness, and meaning-making.
|
🧪 Key Experiments in Encoding
|
Craik & Tulving (1975): Found that words processed semantically were recalled more than those processed visually or acoustically.
Hyde & Jenkins (1969): Participants who judged pleasantness of words (deep processing) recalled more than those who counted letters (shallow).
Bower et al. (1969): Hierarchical organization during encoding improves recall.
Bransford & Johnson (1972): Context helps encoding; participants recalled more when given meaningful context.
|
🧩 Factors Influencing Encoding
|
Attention: Essential for effective encoding—without attention, information decays rapidly.
Elaboration: Linking new information to prior knowledge improves encoding.
Distinctiveness: Unusual or unique items are encoded more deeply.
Rehearsal Type: Elaborative rehearsal (meaning-based) is superior to maintenance rehearsal (rote repetition).
Organizational Strategies: Chunking, imagery, and mnemonics enhance encoding efficiency.
|
🧬 Neuroscience of Encoding
|
Encoding is supported by the hippocampus, prefrontal cortex, and medial temporal lobes.
Hippocampus plays a critical role in consolidating encoded information into long-term memory.
Prefrontal cortex assists in attentional control and selecting encoding strategies.
Neuroimaging (fMRI, PET) shows increased activity in the left hemisphere for verbal encoding, and right for visual encoding.
Neurotransmitters like acetylcholine and glutamate are involved in encoding processes.
|
🧠 Encoding Specificity Principle (Tulving & Thomson, 1973)
|
Recall is most effective when retrieval conditions match encoding conditions.
Context-dependent memory: Environmental cues present during encoding aid retrieval.
State-dependent memory: Internal states (mood, drug-induced states) influence recall.
Mood-congruent memory: We recall information consistent with our current mood.
|
✔️ Practical Applications of Encoding Research
|
Educational psychology: Encouraging meaningful learning and elaboration improves academic performance.
Memory rehabilitation: Techniques like chunking, visualization, and association aid memory-impaired individuals.
Cognitive therapy: Re-encoding traumatic memories in safer, new emotional contexts (e.g., EMDR).
|
❌ Encoding Failures
|
Encoding failure occurs when information never enters long-term memory due to lack of attention or processing.
Common in divided attention tasks or passive learning environments.
Forgetting is often due to ineffective encoding, not memory decay.
|
Retrieval Processes
📌 What is Retrieval?
|
Retrieval refers to the process of accessing stored information from long-term memory.
It is the final stage in the memory process, after encoding and storage.
Retrieval is influenced by how the information was encoded, the type of memory, and retrieval conditions.
Retrieval can be intentional (effortful) or spontaneous (automatic).
|
🧭 Retrieval Cues
|
Retrieval cues are stimuli or triggers that assist in accessing stored memories.
They can be external (environmental, verbal hints) or internal (emotional state, mental associations).
Effective cues often involve associative links formed during encoding.
Cue overload principle: A cue is less effective if it is linked to many items.
Distinctive cues enhance retrieval by reducing interference.
|
🌍 Context-Dependent Retrieval
|
Memory is better retrieved in the same context in which it was encoded.
This includes physical surroundings, people, smells, lighting, and ambient sounds.
Classic study: Godden & Baddeley (1975) found divers recalled more words when encoding and retrieval occurred underwater or both on land.
Context acts as a retrieval scaffold, facilitating access to stored traces.
|
🧠 State-Dependent Retrieval
|
Retrieval improves when a person’s internal physiological or psychological state matches their state during encoding.
Includes effects of mood, arousal, drugs, fatigue, or stress.
Common example: people intoxicated at encoding may recall better when intoxicated again.
Supports the idea that internal states function like retrieval cues.
|
🎭 Mood-Congruent Memory
|
We are more likely to recall memories that match our current emotional state.
This is not about encoding state, but about bias in retrieval content.
Depressed individuals, for example, tend to recall more negative life events.
|
🔄 Recall vs. Recognition
|
Recall: Retrieval without direct cues. Requires reconstructing information.
Examples: Essay tests, free recall tasks.
Types: Free recall, serial recall, and cued recall.
Typically more demanding than recognition.
Recognition: Identifying previously learned information when it is presented again.
Examples: Multiple choice questions, face recognition.
Less effortful—relies on familiarity and retrieval matching.
Recognition is often more accurate than recall due to cue support.
|
🧪 Key Experiments and Theories in Retrieval
|
Tulving’s Encoding Specificity Principle: Retrieval is most effective when cues match the encoding context.
Godden & Baddeley (1975): Environmental context effects in divers.
Eich (1975): Demonstrated state-dependent learning using mood induction.
Loftus (1975): Misinformation effect—shows how retrieval can be distorted by post-event information.
Nelson (1971): Showed that forgotten items can be retrieved when original cues are reinstated.
|
🧬 Neurocognitive Aspects of Retrieval
|
Hippocampus: Essential for relational memory retrieval and reactivating stored memory patterns.
Prefrontal cortex: Involved in retrieval effort, monitoring, and decision-making during recall.
Parietal lobes: Associated with subjective experience of remembering, like familiarity.
Retrieval involves pattern completion: reinstating parts of the stored trace using cues.
|
🔁 Retrieval Practice (Testing Effect)
|
Repeated retrieval strengthens memory more than passive review.
Roediger & Karpicke (2006): Testing enhances long-term retention better than re-studying.
Retrieval promotes reconsolidation and deepens encoding pathways.
|
⚠️ Retrieval Failures and Blocking
|
Retrieval failures are not always due to forgetting—can be caused by:
Interference (retroactive/proactive),
Cue-dependent forgetting,
Decay of the memory trace,
Inhibition or motivated forgetting (e.g., repression).
Tip-of-the-Tongue (TOT) phenomenon: Partial retrieval; activation without full access.
Blocking: Interference from competing memories (e.g., similar names).
|
Forgetting
What is Forgetting? |
orgetting refers to the inability to retrieve information previously encoded and stored in memory.
It may occur due to weak encoding, interrupted consolidation, trace decay, retrieval failure, motivated forgetting, or errors in memory processing.
It’s not always dysfunctional—it helps cognitive efficiency by allowing us to filter irrelevant or outdated information.
|
The Seven Types of Forgetting (Schacter’s “Seven Sins of Memory”) |
a. Transience
Forgetting that occurs with the passage of time.
Memory traces become weaker or degrade if not recalled or rehearsed.
Closely related to trace decay theory.
|
b. Absent-Mindedness
Forgetting due to a lack of attention or shallow encoding.
Often results from distraction or divided attention at the time of encoding.
Example: Forgetting where you placed your keys.
|
c. Blocking
Temporary inability to access stored information.
Often manifests as the Tip-of-the-Tongue (TOT) phenomenon.
Memory is available but inaccessible at that moment.
|
d. Misattribution
Assigning a memory to the wrong source (e.g., thinking someone else told you something).
Can contribute to false memories and distorted recall.
|
e. Suggestibility
Incorporation of misleading information from external sources into personal recollections.
Often observed in eyewitness testimony and memory distortion due to leading questions.
|
f. Bias
Retrospective distortions caused by current beliefs, emotions, or knowledge.
People reshape past events to better fit their present view of themselves or the world.
|
g. Persistence
Unwanted memories that intrude into consciousness.
Often emotionally charged, and seen in PTSD or rumination.
Contrary to typical forgetting – it’s the inability to forget.
|
Trace Decay Theory
|
Suggests that memory traces fade over time if not actively rehearsed.
Based on the physiological decay of memory traces in the brain.
Applies best to sensory memory and short-term memory.
Peterson & Peterson (1959): Demonstrated rapid STM forgetting when rehearsal was blocked.
|
Interference Theory
|
Proposes that conflicting information disrupts memory retrieval.
Two key types:
|
Proactive Interference (PI)
Older memories interfere with the learning or recall of new material.
Example: Using your old PIN when trying to recall a new one.
|
Retroactive Interference (RI)
New information interferes with the retrieval of older memories.
Example: Forgetting your old address after memorizing your current one.
Underwood (1957): Found evidence for PI in list-learning studies.
McGeoch & McDonald (1931): RI is stronger when materials are similar.
|
Motivated Forgetting (Freudian Theory)
|
Originates from Freud’s psychodynamic theory.
Proposes that people forget emotionally disturbing or threatening memories to protect the ego.
Two main forms:
Repression: Unconscious blocking of distressing memories.
Suppression: Conscious, intentional effort to avoid remembering.
Anderson & Green (2001): Experimental support via Think/No-Think paradigm.
|
Retrieval Failure (Cue-Dependent Forgetting)
|
Memory is stored but cannot be accessed due to a lack of proper retrieval cues.
Explained by Encoding Specificity Principle (Tulving): retrieval is most effective when context matches encoding.
Examples: Forgetting a name until reminded by a mutual friend.
|
Additional Concepts Related to Forgetting
|
Tip-of-the-Tongue (TOT) Phenomenon
Partial retrieval failure – the feeling of knowing something but being unable to retrieve it.
b. Consolidation Failure
Forgetting due to interruption or failure during memory consolidation, often due to trauma or interference.
c. Directed Forgetting
Intentional forgetting due to instructions or cognitive control.
Studied using item-method and list-method paradigms.
d. Organic Causes of Forgetting
Brain damage, neurodegenerative diseases (e.g., Alzheimer’s, Korsakoff’s syndrome), and trauma can impair memory.
These typically affect episodic and semantic memory, but procedural memory often remains intact.
|
Neurobiological Aspects of Forgetting
|
Hippocampus: Crucial for memory consolidation; damage results in anterograde or retrograde amnesia.
Prefrontal Cortex: Involved in retrieval, inhibition of unwanted memories, and cognitive control.
Forgetting may also result from synaptic pruning and long-term depression (LTD) – reduction in synaptic strength.
Neurotransmitters like glutamate, GABA, and acetylcholine influence memory encoding and stability.
|
|
