Notes on Sensory Memory and the Modal Model of Memory

Sensory Memory and the Modal Model of Memory

• The big theme: how memory is structured and how sensory input becomes memory. The teacher uses a modal model to orient discussion: sensory store/register → short-term memory (STM) → long-term memory (LTM). The term “modal model” reflects common family of similar models, not a single author. The boxes are used as schematic representations of memory stores; note the size mismatch in visual slides (STM and LTM are not the same size in reality).

The Modal Model of Memory

• Information from the environment first hits a sensory store/register (sensory memory).

• It then goes to short-term memory (STM).

• With attention and rehearsal, it may transfer to long-term memory (LTM).

• Practical note: STM capacity is small; the classic view cites about 7 \pm 2 items, but more recent work suggests the effective raw capacity without chunking is closer to 4 \pm 1 items.

• Duration varies by modality and can be extended by maintenance rehearsal.

• Working memory is the modern rebranding often used for STM in task contexts, but this lecture sticks with short-term memory terminology for now.

• Key historical figures and ideas:

  • Miller (1956) introduced the concept of the “magic number” 7 \pm 2 items for short-term memory capacity.
  • Cowan and colleagues have argued for a more constrained raw capacity around 4 \pm 1 without chunking or mnemonic strategies.
  • Chunking (recoding) allows people to remember more by grouping items into meaningful units using long-term memory to bootstrap short-term storage.

• Everyday test logic: memory tests sample a subset of material (e.g., a 7-item list) to infer overall memory ability, rather than testing every item.

Visual Sensory Memory (Iconic Memory)

• Definition: a temporary visual buffer that supports continuity of perception across rapid eye movements (fixations) and prevents a jagged visual world.

• Visual persistence helps the visual system form a continuous experience despite saccades and brief fixations.

• Capacity and duration:

  • Visual capacity is large for a brief instant; you can hold about 75\% of a visual scene for about 0.25\ \text{s} (a quarter of a second).
  • Afterimage and color retention demonstrate that this memory is not purely retinal; there is a memory component.

• Classic research hook: the partial report method (Spurling): briefly flashed array (approx. 250\ \text{ms}) with the later report cued by a tone to a specific row (top/middle/bottom). This demonstrated that observers could access about 3 of 4 items in a row, implying the entire scene was available briefly, but decay/limitations reduced reportability.

• Whole report vs partial report:

  • Whole report: participants were asked to report everything they saw, which drastically underestimates capacity because items fade before reporting.
  • Partial report: cue after stimulus (tone indicating which row to report) yields much higher accuracy for that row, revealing high initial visual storage capacity.

• Methodological impact: this clever design bypassed the limitation of the whole report, showing that people can see far more than they can report at once, given the right cue.

• Historical note: the visual sensory memory field flourished in the 1960s–70s, but later faced critiques about ecological validity and variability across individuals. The field is not a major focus in modern practice, though the phenomenon remains a classic demonstration of sensory memory.

• Visual masking and afterimages: masking can suppress or obscure information in the visual channel, similarly to auditory masking, indicating interactions between memory and ongoing perception.

• Practical note: iconic memory is foundational but not highly used in isolation today; the auditory sensory memory is often considered more ecologically valid for understanding everyday cognition (speech, sounds).

• Summary points for visual sensory memory:

  • Capacity: high for a brief moment; about 75\% of a scene accessible in the instant after viewing.
  • Duration: ≈ 0.25\ \text{s} without masking.
  • Coding: largely raw sensory in this stage; meaning and interpretation come later in processing.

• Visual vs auditory memory (ecological validity): visual memory studies tended to lack ecological validity; auditory memory studies tend to map more directly onto real-world perception (speech, environmental sounds).

Auditory Sensory Memory

• Auditory memory lasts longer than visual memory, with estimates around 3 \text{--} 4\ \text{s} without masking or distraction.

• Auditory domain is inherently temporal and crucial for speech perception and music; the brain must hold onto a spoken sentence long enough to process it.

• Demonstrations and concepts are often used to illustrate how auditory memory supports perception and language processing.

• Modality effect: there is a robust finding that the recency effect (better memory for last items) is stronger in auditory lists than in visual lists. This is attributed to the persistence of auditory sensory memory for a few seconds after the items stop, allowing last items to be recalled more readily if not masked.

• Demonstration: a class activity comparing auditory vs visual list recall shows stronger recency for the auditory list, consistent with the modality effect.

• Masking and suffix effects in auditory memory:

  • Masking: additional sounds can mask the memory trace of the target stimulus, reducing recall.
  • Suffix effect: adding a suffix (e.g., the word “end”) after the list eliminates the modality effect and collapses auditory and visual curves; essentially, the post-list signal overwrites or displaces the last items in auditory STM.
  • Practical implication: silent pauses can boost retention of auditory items, but adding a cue that overlaps with the memory trace can erase the advantage.

• Auditory memory in real-world tasks: examples include two-factor authentication workflows (reading codes aloud or typing after hearing them), where auditory encoding can aid memory performance.

• Summary points for auditory sensory memory:

  • Duration: ≈ 3 \text{--} 4\ \text{s}.
  • Recency advantage (modality effect) is more pronounced than in the visual domain.
  • Masking and suffix effects illustrate the vulnerability and dynamics of auditory trace maintenance.

Partial Report, Whole Report and the Spurling Legacy

• Spurling’s partial report method (Harvard origin) demonstrated a clever approach to measuring sensory memory capacity without the confounds of immediate reporting.

• Parsimonious design: brief presentation (≈ 250\ \text{ms}), followed by a cue that directs which subset to report; participants reliably report the cued subset, implying the entire array was available in sensory memory but was not all reportable at once.

• This line of work heavily influenced subsequent memory research and methodological design for decades.

• Takeaways:

  • Sensory memory has substantial capacity for a brief moment, but the information decays rapidly unless attention and rehearsal occur.
  • The modality (visual vs auditory) matters for the duration and the presence of recency effects.

Serial Position Curve: Primacy and Recency

• Serial position curve: plotting recall probability as a function of an item’s position in a list (first to last).

• Primacy effect: better recall for items at the beginning of the list, attributed to more attention, rehearsal, and possible transfer to LTM.

  • Developmental note: primacy effects emerge with schooling; younger preschool children show less primacy because they have had fewer opportunities to rehearse and chunk information.

• Recency effect: better recall for items at the end of the list; strongly linked to short-term memory and sensory memory traces.

• Modality effect in recency: stronger for auditory lists than visual lists due to the persistence of auditory memory for a few seconds after list termination.

• Suffix effect and other manipulations can erase the recency advantage, especially in auditory memory, by introducing interference at the end of the list.

• Practical implications:

  • Immediate testing tends to show strong recency effects, particularly in the auditory domain, because items remain within the sensory/short-term buffer.
  • Delayed testing or interference can shift the balance toward primacy (long-term encoding of early items) or erase recency advantages.

Short-Term Memory (STM) and Working Memory

• Short-term memory is the buffer that holds information for a short period (seconds) unless rehearsed.

• Capacity debates:

  • Classic view: around 7 \pm 2 items (Miller, 1956).
  • Contemporary view with minimal chunking: around 4 \pm 1 items (Nelson Cowan and colleagues).

• Duration without rehearsal: typically only a few seconds; with maintenance rehearsal, information can be kept longer and potentially transferred to LTM.

• The concept of chunking: grouping smaller units into meaningful units to increase effective capacity. For example, turning a string such as 13902 into a meaningful chunk (e.g., a ZIP code) reduces the number of items tracked in short-term memory.

• Classic experiments and notable narratives:

  • Miller’s magic number (7 ± 2) – often cited as the capacity of STM without chunking.
  • Chunking and recoding strategies shown to dramatically increase effective STM capacity when people use long-term memory to organize information.
  • Ericsson and Chase explored extreme chunking abilities with trained individuals; one participant achieved extraordinary digit spans by using elaborate mnemonic schemes (cross-country running-time style chunking). Even with intensive training, some individuals still show limits when chunking is constrained by the same stimulus set.
  • The famous “SF” participant demonstrated far higher digit spans with extended chunking practices, illustrating how long-term memory and structured encoding strategies can temporarily boost STM-like performance for specific tasks.

• The key takeaway: STM capacity is not fixed; chunking and mnemonic strategies can dramatically expand what seems possible within the short-term buffer. However, these gains are often highly task- and stimulus-specific and do not generalize to arbitrary material (e.g., unrelated letters).

• Practical implications:

  • In real-world tasks, speakers/readers often rely on chunking to hold sequences (phone numbers, codes, etc.).
  • The proper use of encoding strategies can meaningfully increase performance in memory tasks, especially when the content is chunkable or has meaningful structure.

• The “magic” number debate culminates in Cowan’s view that, without chunking or domain-specific strategies, the raw capacity is closer to 4 \pm 1 items, not the traditional 7 \pm 2.

• A fun aside on subitizing: up to four units are quickly counted without explicit enumeration; five or more require rapid counting. This aligns with a rough upper bound around four items for quick apprehension without specialized chunking.

Notable Studies, People, and Takeaways

• Spurling partial-report methodology demonstrated access to a large but fleeting visual store.
• Miller (1956): introduced the idea of seven items as a typical STM capacity, popularized as the “magic number.”
• Cowan and colleagues (modern view): raw capacity closer to 4 \pm 1; chunking is essential to achieving higher performance on many tasks.
• Chase and Ericsson: elite memory performance via structured chunking strategies; demonstrates interaction between STM and LTM and the power of encoding techniques.
• Subitizing: fast recognition of small numbers (roughly up to 4) without counting, reflecting limits in raw capacity.

Practical Implications for Studying and Exams

• Expect questions about:
  • The structure of the modal model (sensory memory, STM, LTM) and why we call it a model, not a single theory.
  • Differences between iconic (visual) and echoic (auditory) memory in duration, capacity, and ecological validity.
  • The partial-report method and what it reveals about sensory memory capacity.
  • The modality effect and the suffix effect, and what they imply about how memory is refreshed or displaced by new information.
  • Serial position effect and the distinct explanations for primacy vs recency.
  • Short-term memory capacity theories (7 ± 2 vs 4 ± 1) and the role of chunking.
  • The tight link between short-term memory and long-term memory; how chunking can involve long-term memory to aid short-term retention.
  • Real-world applications of memory principles (e.g., interface design, authentication processes, education strategies).

Quick Reference Formulas and Key Numbers

• STM capacity (classic): 7\pm 2 items

• Raw STM capacity (modern view, with minimal chunking): 4\pm 1 items

• Visual iconic memory duration: \approx 0.25\ \text{s}

• Visual memory usable fraction of scene: \approx 75\%

• Visual memory capacity (brief): up to nearly entire scene, but decay/displacement reduce reportability

• Auditory sensory memory duration: \approx 3\text{--}4\ \text{s}

• Temporal storage acts to maintain speech and auditory streams beyond the moment of articulation

• Spurling partial-report cue timing: stimulus duration \approx 250\ \text{ms}; cue follows to reveal subset recall

• Serial position effects: primacy (beginning items) and recency (end items)

• Chunking: recoding into meaningful units; capacity can be dramatically expanded by organizing information

• Subitizing limit: roughly up to 4 units for rapid perception without counting

• Note: When you see numbers in these notes, you’ll often see them presented in LaTeX form, such as 7\pm 2, 4\pm 1, or durations like 0.25\ \text{s} to keep mathematical precision consistent with the course style.

• Final takeaway: sensory memory provides the raw, brief foothold for perception; short-term memory handles immediate recall and manipulation; long-term memory stores durable knowledge. The capacity and duration of these stores depend on modality and strategies like chunking and rehearsal, which can dramatically change what we can hold in mind at one time.