Premed · Premed · Introductory Psychology
Lecture 11: Memory: Encoding, Storage, Retrieval
Introductory Psychology
Learning Objectives
By the end of this lecture, students will be able to:
- Define memory and describe the three stages of memory processing
- Explain the Atkinson-Shiffrin model and the working memory model
- Distinguish between explicit and implicit memory systems
- Describe effective encoding strategies including levels of processing and elaborative rehearsal
- Explain the biological basis of memory including long-term potentiation
Lecture Content
I. What Is Memory?
Memory is the persistence of learning over time through the encoding, storage, and retrieval of information. Encoding is the process of getting information into memory by transforming sensory input into a neural code. Storage is the retention of encoded information over time. Retrieval is the process of accessing stored information when it is needed. An essential point is that memory is not a perfect recording of experience — it is a constructive and reconstructive process that can introduce errors and distortions.
II. The Atkinson-Shiffrin Model (Multi-Store Model, 1968)
The Atkinson-Shiffrin model proposes three distinct memory stores, each with different capacities and durations. Sensory memory holds raw sensory information for a very brief period. Iconic memory (visual) lasts roughly half a second; George Sperling's 1960 experiments with whole-report versus partial-report procedures demonstrated that iconic memory has a large capacity but fades almost instantly. Echoic memory (auditory) persists for about three to four seconds. The function of sensory memory is to hold incoming sensory data just long enough for selective attention to pick out what matters.
Short-term memory (STM) has a limited capacity of approximately seven plus or minus two items, as George Miller described in his classic 1956 paper "The Magical Number Seven." Information in STM lasts about 15 to 30 seconds without rehearsal. Maintenance rehearsal — simply repeating information — keeps it active in STM but represents shallow processing. Chunking, the strategy of grouping items into meaningful units, can increase the effective capacity of STM: the string FBICIAUSANBA is hard to remember as 12 individual letters but easy to recall as four familiar chunks (FBI, CIA, USA, NBA).
Long-term memory (LTM) has essentially unlimited capacity and duration. Information transfers from STM to LTM through encoding processes such as rehearsal, elaboration, and organization.
III. Working Memory Model (Baddeley, 2000)
Alan Baddeley's working memory model updated the concept of short-term memory by emphasizing active processing rather than mere passive storage. The model includes four components. The central executive is the attentional control system that directs focus, coordinates information from the other components, and manages cognitive resources — functioning as the "boss" of working memory with limited capacity. The phonological loop processes and rehearses verbal and acoustic information through a phonological store (which holds sound-based representations) and an articulatory rehearsal process (the "inner voice"). The word-length effect — the finding that long words are harder to remember because they take longer to rehearse — provides evidence for this loop. The visuospatial sketchpad processes and manipulates visual and spatial information, supporting mental imagery, spatial reasoning, and navigation. The episodic buffer, added by Baddeley in 2000, integrates information from the other components and from long-term memory into coherent episodes. Working memory capacity is a strong predictor of intelligence, reading comprehension, and academic performance.
<image>A side-by-side comparison of two memory models. Panel A: The Atkinson-Shiffrin model shown as three boxes connected by arrows: Sensory Memory (very large capacity, very brief duration) → attention filter → Short-Term Memory (7 +/- 2 items, 15-30 seconds) → encoding → Long-Term Memory (unlimited capacity, potentially permanent). A rehearsal loop circles back within STM. A retrieval arrow points back from LTM to STM. Panel B: Baddeley's Working Memory model shown as four components: a central executive at the top connected to three subsystems below — the phonological loop (with ear icon), the visuospatial sketchpad (with eye icon), and the episodic buffer (connecting to long-term memory). Arrows show bidirectional information flow.</image>
IV. Long-Term Memory Systems
Long-term memory is divided into two broad categories. Explicit (declarative) memory consists of memories that can be consciously recalled and verbalized. It further divides into episodic memory — personal experiences and events such as "what I had for breakfast" or "my first day of college," rich with contextual details of time, place, and emotion — and semantic memory, which stores general knowledge and facts about the world, such as "Ottawa is the capital of Canada," without being tied to any particular personal experience.
Implicit (nondeclarative) memory influences behavior without conscious awareness. Procedural memory stores skills and habits such as riding a bike, typing, or playing an instrument, and involves the cerebellum and basal ganglia. Classical conditioning effects, including conditioned emotional responses and conditioned reflexes, rely on the amygdala (for emotional conditioning) and cerebellum (for motor conditioning). Priming occurs when prior exposure to a stimulus facilitates later processing of the same or a related stimulus — seeing the word "doctor," for example, speeds up recognition of the word "nurse."
The hippocampus is critical for encoding new explicit memories through a process of consolidation. Patient H.M. (Henry Molaison), who underwent bilateral hippocampal removal, developed severe anterograde amnesia for explicit memories while his implicit memory remained intact, providing some of the most compelling evidence for the distinction between these memory systems. The amygdala tags memories with emotional significance, the cerebellum and basal ganglia support procedural learning, and the prefrontal cortex is involved in working memory and retrieval strategies.
V. Encoding: Getting Information into Memory
The levels-of-processing framework, proposed by Craik and Lockhart in 1972, argues that memory depends on the depth at which information is processed, not merely on how long it is held in short-term memory. Shallow processing involves attending to structural or surface-level features ("Is the word in uppercase?"). Intermediate processing involves phonemic or acoustic encoding ("Does the word rhyme with 'cat'?"). Deep processing involves semantic encoding — engaging with meaning ("Does the word fit in this sentence?"). Deeper processing consistently produces stronger, more durable memories.
Elaborative rehearsal, which involves linking new information to existing knowledge and personal experience, is far more effective than maintenance rehearsal for long-term retention. The self-reference effect demonstrates that information related to oneself is encoded more deeply and remembered more easily. The dual coding theory, proposed by Paivio, suggests that encoding information both verbally and visually creates two retrieval routes, which is why concrete words (like "apple" or "dog") are easier to remember than abstract ones (like "justice" or "democracy").
Organization improves encoding in several ways. Hierarchical organization arranges information in categories and subcategories. Schemas — existing mental frameworks — help organize and interpret new information. Mnemonic devices such as the method of loci (associating items with locations along a familiar route) and the peg-word system (associating items with a pre-memorized list of words) leverage organizational structures to aid recall.
Among the most robust findings in memory research is the spacing effect: distributing study sessions over time (distributed practice) produces far better retention than cramming everything into one session (massed practice). The testing effect, also called retrieval practice, shows that actively retrieving information from memory strengthens it more effectively than simply re-reading the material. Interleaving — mixing different types of problems or topics during study — further enhances learning.
VI. Retrieval: Getting Information Out
Retrieval cues are stimuli that help access stored memories. Recall involves retrieving information with minimal cues, as in a fill-in-the-blank question. Recognition involves identifying previously learned information from a set of options, as in a multiple-choice question, and is typically easier because more cues are available. Relearning, also known as the savings method, was pioneered by Ebbinghaus and demonstrates that material learned once can be relearned faster the second time — evidence that some retention persists even when conscious recall fails.
The encoding specificity principle, formulated by Tulving, states that memory is best when the conditions at retrieval match those at encoding. Context-dependent memory reflects this principle: the classic scuba diver study by Godden and Baddeley (1975) showed that words learned underwater were better recalled underwater. State-dependent memory extends the principle to internal states: memory is enhanced when the person's mood or physiological state during retrieval matches their state during encoding. Mood-congruent memory means that our current mood influences which memories are most accessible — when sad, we more readily recall sad memories, a pattern that can feed into depression.
<image>A hierarchical diagram of long-term memory systems. At the top, "Long-Term Memory" branches into two main categories: "Explicit (Declarative)" and "Implicit (Nondeclarative)." Explicit branches into "Episodic" (with an icon of a birthday party photo) and "Semantic" (with an icon of a textbook). Implicit branches into "Procedural" (icon of hands on a piano), "Classical Conditioning" (icon of Pavlov's dog), and "Priming" (icon of word fragments completing). Below each type, the primary brain region is noted: hippocampus for explicit, cerebellum and basal ganglia for procedural, amygdala for conditioning, and cortex for priming.</image>
VII. Biological Basis of Memory
Long-term potentiation (LTP) is the prolonged strengthening of synaptic connections following repeated stimulation. It primarily involves glutamate and NMDA receptors in the hippocampus and is believed to be the fundamental neural mechanism underlying learning and memory formation. At the synaptic level, learning involves increased neurotransmitter release, a greater number of receptors on the postsynaptic membrane, and the growth of new dendritic spines and synaptic connections.
Memory consolidation is the process by which fragile short-term memories are converted into stable long-term ones. This process unfolds over hours to weeks and is strongly facilitated by sleep — particularly slow-wave sleep and REM sleep. During sleep, the hippocampus replays recent experiences, effectively transferring information to the cortex for more permanent storage. Reconsolidation is a more recently discovered process in which a retrieved memory becomes temporarily destabilized and must be re-stored. This finding has significant implications because it means memories can potentially be modified during reconsolidation, opening the door to therapeutic applications for conditions such as PTSD.
