Major Models Of Memory

Multistore Model

Atkinson & Shiffrin (1968)

• Memory has 3 stores:
  • Sensory store → briefly holds sensory information
  • Short-term memory (STM) → temporary, limited-capacity storage
  • Long-term memory (LTM) → long-term, very large capacity
• Information moves:
  • Sensory store → STM → LTM
• Rehearsal was proposed to transfer information from STM → LTM.

Problems with the Multistore Model

• Too focused on separate storage structures.
• STM and LTM are probably not single, unitary stores.
• Rehearsal does not always lead to long-term learning.
• LTM can influence STM.
  • e.g. CIATVHSC → CIA / TV / HSC
  • Existing knowledge allows information to be chunked.
• STM may not be necessary for information to enter LTM.

Patient KF

• Brain injury caused severely impaired STM.
• Memory span ≈ 1 item.
• However, LTM was relatively intact.
• Could learn word pairs and remember them after 24 hours.
• Suggests:
  • STM and LTM are separable.
  • Information does not necessarily have to pass through an intact STM system to reach LTM.

STM vs Working Memory

Short-Term Memory

• Temporary storage of information.
• Limited capacity.
• Information easily lost without rehearsal.

Working Memory

• Temporarily stores AND manipulates information.
• Used during active thinking.
• e.g.
  • Mental arithmetic
  • Reading
  • Problem solving
  • Following instructions
• Made up of multiple specialised components.

Baddeley’s Working Memory Model

Four main components:

• Central Executive (CE)
  • Controls attention.
  • Coordinates other components.
  • Allocates mental resources.
• Phonological Loop (PL)
  • Verbal + auditory information.
• Visuospatial Sketchpad (VSSP)
  • Visual + spatial information.
• Episodic Buffer (EB)
  • Combines information from different sources.
  • Links working memory with LTM.

Easy memory trick:

• PL = hear/say
• VSSP = see/where
• EB = bind
• CE = control

Phonological Loop

• Stores verbal/auditory information.
• Holds about 1.5–2 seconds of spoken material.

Two Parts

Phonological store

• Temporarily stores speech-based information.
• Passive storage.

Articulatory rehearsal process

• Mentally repeats information.
• Refreshes information before it disappears.
• Like your inner voice.

Phonological Similarity Effect

• Similar-sounding words are harder to remember in the correct order.
• e.g. words like:
  • cat
  • mat
  • bat
  • rat
• More easily confused than words that sound different.

Why?

• Verbal information is stored using sound-based/phonological codes.
• Similar sounds create more confusion.

Supports

• Existence of the phonological loop.
• Especially the phonological store.

Articulatory Suppression

• Participant repeatedly says an irrelevant sound/word while trying to remember information.
• e.g. repeating “blah blah blah…”

Effect

• Prevents normal verbal rehearsal.
• Recall becomes worse.

Why?

• The articulatory rehearsal process is already occupied.
• Person cannot keep refreshing the target information.

Therefore:

• Rehearsal is important for maintaining verbal information in working memory.

Word Length Effect

Baddeley et al. (1975)

• Short words are remembered better than long words.

Example:

• Short: bond, wit, harm
• Long: individual, opportunity, representative

Why?

• Working memory appears limited partly by time.
• We can hold around 1.5–2 seconds of verbal information.
• Short words can be rehearsed faster.
• More short words can therefore be refreshed before they decay.

Important

• It is mainly about how long the word takes to say, not simply number of syllables.
• e.g. harpoon can be harder than bishop if it takes longer to articulate.

Digit Span Across Languages

• Digit span differs between languages.
• Lecture pattern:
  • Chinese > English > Welsh

Explanation

• Digit names take different amounts of time to pronounce.
• Shorter/faster digit names → more digits can be rehearsed within the phonological loop.

Articulatory Suppression

• When participants cannot rehearse:
  • Digit span decreases.
  • Differences between languages become smaller.
• Supports role of rehearsal speed.

BUT

• Chinese advantage does not disappear completely.
• Suggests other language-specific factors may also contribute.

Phonological Loop & Language Learning

Patient PV

• Very poor verbal STM/digit span.
• Could learn associations involving familiar/native words.
• Struggled to learn new Russian words.

Suggests

• Phonological loop is especially important when learning unfamiliar sound patterns.
• Important for new vocabulary acquisition.

Healthy Participants

• Articulatory suppression particularly disrupts learning of foreign vocabulary.
• Less disruptive for associations between already familiar words.

Therefore:

• Phonological loop helps temporarily maintain unfamiliar sounds while they are being learned.

Visuospatial Sketchpad

• Stores/manipulates visual and spatial information.
• Used for:
  • Mental imagery
  • Navigation
  • Remembering locations
  • Mental rotation
  • Faces
  • Scenes/layouts

Two Components

Visual Cache

• Stores visual appearance.
• e.g.
  • Shape
  • Colour
  • Visual details

Inner Scribe

• Stores spatial/movement information.
• e.g.
  • Locations
  • Movement
  • Spatial sequences

Episodic Buffer

• Temporary, limited-capacity workspace.
• Combines information from different sources.
• Integrates:
  • Verbal information
  • Visual information
  • Spatial information
  • LTM knowledge

Function

• Creates a single, coherent episode.

Example:

• Hearing a phone number while seeing its position on a keypad.
• Episodic buffer can bind the verbal digits + spatial locations.

Visuospatial Bootstrapping

• Memory for verbal information can improve when it is paired with useful spatial information.
• e.g. remembering digits using a familiar keypad layout.

Why?

• Uses more than one working-memory system.
• Verbal information → phonological loop.
• Spatial layout → visuospatial sketchpad.
• Episodic buffer → binds them together.

Evidence for Separate Working Memory Components

Dual-Task Studies

• People can often perform two tasks reasonably well when they use different working-memory components.

Example:

• Repeating words → phonological loop.
• Tracking a moving object → visuospatial sketchpad.

→ Relatively little interference.

BUT:

• Two tasks using the same component interfere more strongly.

Conclusion

• Working memory is not one general storage system.
• Contains specialised, relatively independent components.

Central Executive

• Controls and coordinates working memory.
• Modality-free → not specific to visual or verbal information.
• Limited capacity.

Functions

• Controls attention.
• Focuses on task goals.
• Suppresses distractions.
• Coordinates other working-memory systems.
• Switches between tasks.
• Plans/sequences behaviour.
• Manages competing tasks.

Limited Capacity

• Too many demanding tasks → slower performance + more errors.

Executive Functions

Higher-level cognitive control processes.

Main examples:

• Inhibition
  • Suppress an automatic/irrelevant response.
• Shifting
  • Switch between tasks or mental sets.
• Updating
  • Replace old information in working memory with currently relevant information.

Norman & Shallice Model

Schemas

• Learned/automatic action routines.
• Triggered by environmental cues.

Examples:

• Red traffic light → stop.
• Seeing a written word → automatically read it.

Contention Scheduling

• Controls routine/automatic behaviour.
• Selects between familiar schemas.
• Basically = autopilot.

Supervisory Attentional System (SAS)

• Used when automatic behaviour is not enough.
• Provides conscious, flexible control.
• Important for:
  • Novel situations
  • Difficult tasks
  • Conflicting responses
  • Overriding habits

Easy distinction:

• Contention scheduling = autopilot
• SAS = deliberate control when autopilot isn’t appropriate

Stroop Task

Example:

RED written in blue ink.

Task = say “blue.”

Problem

• Reading the word RED is highly automatic.
• Need to inhibit word reading and focus on ink colour.

Requires

• Inhibition
• Central executive/SAS control

Stroop Effect

• Slower/more errors when:
  • Word meaning and ink colour conflict.

Shows

• Automatic processes can compete with current goals.
• Executive control is needed to suppress the automatic response.

Dysexecutive Syndrome

• Impairment of executive control.
• Often associated with prefrontal cortex damage.

Can involve difficulty with:

• Planning
• Inhibition
• Flexible behaviour
• Task switching
• Maintaining goals
• Controlling automatic responses