Memory

Phases

Acquisition
Storage
Retrieval

Information processing modal model

Sensory information
-> sensory memory: raw form, brief (iconic = visual, echoic = auditory)
-> short-term memory: if attended, maintenance via rehearsal
-> long-term memory

Working vs long-term memory

Working memory

Status, not place
Thoughts currently active
Needs rehearsal to maintain
Used for multiple ideas - combine & compare
Size: limited (9-10 items for 30s)
Ease of entry & retreival: easy
Fragility of contents: fragile

Long-term memory

Thoughts not currently active
Stores knowledge, beliefs, episodic (events)
Size: unlimited
Ease of entry & retrieval: hard
Fragility of contents: not fragile

Serial position effect

Serial position effect/curve

U-shaped curve of serial position vs recall likelihood
Shows 2 separate memories

Primacy effect: remember first words - more attention/rehearsal -> LTM

Recency effect: remember last words - still in WM -> WM

Unfilled delay: no effect
Filled delay: eliminates recency, doesn't affect primacy

Word list slower: improves pre-recency, doesn't affect recency
Word list faster: reduces pre-recency, doesn't affect recency

Working Memory

Measuring WM

Digital span task

Repeat increasingly longer series of digits (different each time)
Digit span: no. digits repeated without error
=> capacity = 7 plus-or-minus-2 chunks
Chunking can increase items remembered

Operation span task

Work while retaining something in memory
e.g. T/F maths, then recall word
Measures WM capacity (WMC) - efficiency when working
Larger WMC -> better at coordinating info (reason, read, comprehend)

Cultural differences

Chinese > English in numerical cognition (maths)
Faster at maths OR language differences

Components

Central executive: control centre, coordinates, does 'work'

Visuospatial buffer / sketchpad: stores visual/spatial <-> visual semantics in LTM

Episodic buffer: organises info chronologically <-> episodic in LTM

Phonological (articulatory rehearsal) loop: stores aural/speech-based <-> language in LTM

Subvocalisation: silent speech by 'inner voice' starts loop
Phonological buffer: passive storage system, image in 'inner ear'
Fades after a few seconds
Make sound-alike errors, not visual errors (even for visual info)
If deaf, have similar 'inner hand'
Concurrent articulation task: say 'tah', cannot use rehearsal loop -> memory span reduced from 7-5, no sound-alike errors - across cultures

Entering LTM

Rehearsal types

Maintenance: repeat, little thought -> low recall

Relational / elaborative: seek meaning & relationships -> high recall

Need active encoding (pay attention, think about) to remember

Intentional learning: intention adds nothing to ability to learn

But engagement matters
Intentional learning: deliberate, expectation of test
Incidental learning: no intention, no expectation of test
Hyde & Jenkins: remember as many words in 'How pleasant?' group as 'Find the e' & 'Count the letters' group

Levels of processing

Maintenance < shallow < deep < deep elaborative

Shallow: attention to sound/visual form (superficial) -> poor memory
Deep: attention to meaning -> better memory
Elaborative: in context, rich relationships -> better memory

Connections help retrieval

Deep -> create connections -> more retrieval paths -> better retrieval
Elaborative -> potential for connections -> ...
Better understanding -> many connections -> ...

Mnemonics

Techniques to improve memory by imposing organisation

e.g. acronyms, mental pictures, method of loci, peg-word system
Creates connections -> retrieval paths -> ...

Other cultures

Aboriginal & Torres Strait Islander song lines
Greek/Roman memory palaces / method of loci

Advantages: help memorise (good if meaning not important)
Disadvantages: less understanding & connections

Learning - context

Network - spreading activation

Memory types

Processing fluency
& familiarity

Amnesia

Memory errors

False memories

Forgetting

Autobiographical memory

Very long-term remembering

Learning: creates retrieval paths
Remembering: depends on retrieval cues, uses retrieval paths

Context-dependent learning

Better recall if in same physical place as learning
Because can use connections established earlier
=> context reinstatement: recreate same mental & emotional state as learning - same starting point

Encoding specificity

Learn Information within context - connections change what you remember
Culture: collectivist = holistic (influenced by background), individualist = analytical (not influenced by background)

Memory is network of ideas

Nodes connected via associations

Spreading activation

Node receives activation
-> activation level increases
-> if reaches response threshold
-> node fires
-> activates all connected nodes

Summation of sub-threshold activities, e.g. hints
Can shut down spread
Can control starting points

Semantic priming experiment

Lexical decision: say if 2 strings are words
If related -> semantic priming -> faster
If unrelated -> slower
Because related concepts linked & can activate each other
=> evidence of spreading activation

Types

Retrograde: can't remember experiences before disruptive event
Anterograde: can't remember experiences after disruptive event

Anterograde amnesia: impaired explicit, intact implicit memory

e.g. Shake patient's hand with pin -> patient refuses to shake hand again
e.g. Patient doesn't know trivia answer -> gets it right when asked again but doesn't know why

Double dissociation: 2 types of memory

e.g. blue light -> horn, other lights -> no horn
Hippocampus damage: fear reaction to blue, but couldn't recall blue -> horn => implicit OK, explicit impaired
Amygdala damage: no fear reaction to blue, but could recall blue -> horn => explicit OK, implicit impaired

Recall vs recognition

Recognition

Decide which item from list was presented earlier
Use source memory OR familiarity & attribution
e.g. which did you have for breakfast: toast or cornflakes?
Familiarity: feeling that stimulus was encountered or influence* of stimulus
Know judgements - anterior parahippocampus

Recall

Retrieve from memory in response to recall cue
Use source memory
e.g. what did you have for breakfast?
Source memory: recollection of event, time/place
Remember judgements - hippocampus

Source memory vs familiarity

Independent (can have one without the other
Use remember/know judgements

Implicit vs explicit

Explicit memory

Direct testing
Recall & recognition tests

Implicit memory

Indirect testing (often priming)
Influence with no awareness
Lexical-decision task: see words; decide if strings are real words; if seen before -> faster decision (repetition priming)
Word-stem completion: see words; produce word starting with given 3-4 letters -> produce primed word

False fame effect

False fame: pronunciation list of names, new list of names - rate how famous
Immediate: familiarity + source memory - don't incorrectly believe famous
24 hours later: familiarity but no source memory - incorrectly believe famous
=> misattribute familiarity to fame because familiarity open to interpretation

Illusion of truth

Claims that seem familiar (implicit memory) seem more plausible
e.g. judge how interesting statements are; later, judge how plausible statements are; if heard before -> familiar, so more plausible
e.g. see staged crime, see mugshots of innocents -> 29% falsely select in lineup
misattribute familiarity to credibility - incorrectly judge source

Processing pathway: nodes & connections that activation flows through when recognising
Processing fluency: speed & ease of activation flow - Increased by use

Familiarity as conclusion

Feels unfamiliar

Encounter with stimulus increases processing fluency
-> increased fluency
stop here => unfamiliar

Familiar

Encounter with stimulus
-> increased fluency (or more/less fluency than expected)
-> stimulus feels distinctive
-> attribute distinctiveness to event => familiar

Illusion of familiarity

Manipulation to make perceiving easier
-> increased fluency
-> stimulus feels distinctive
-> attribute distinctiveness to event => familiar

Types

Episodic: specific events
Semantic: general knowledge

Types

Procedural: skills
Priming
Perceptual learning: distinguish sounds/smells
Classical conditioning
Have different functional & biological underpinnings

Make memory errors: even if significant & recent

e.g. Crombag: 50% though saw Dutch plane crash video - significant
e.g. Brewer & Treyens: incorrectly recall books in room - recent

Understanding helps and hurts memory

Help: understanding -> connections -> retrieval path -> locate memory
Hurt: understanding -> connections -> memories interwoven -> 'transplant' info to different context => intrusion error
e.g. story about Nancy & professor with and without context - with context, recall more but make more intrusion errors

DRM effect

See / hear words related to theme, but no theme word -> people 'remember' theme word (even if warned)

Use schemas to fill memory gaps -> errors regularise gaps

Misinformation effect: misinformation effects memory

Memory errors from misinformation received after event, especially if plausible, supplements memory, imagined (not just heard about)
Can't distinguish from false, recalled with as much detail/emotion/confidence
Even memory of self, large-scale, completely false, emotional events - e.g. being lost, hot air balloon ride (T/F photo), slime on teacher's desk
Imagining, paraphrasing, explaining -> false childhood memories - imagination inflation (more confident about imagined > non-imagined events)

Memory confidence poor indicator of accuracy

Confidence reliable in narrow set of circumstances
Because confidence influenced by factors that have no impact on accuracy, e.g. +ve feedback

3 types - all correct

1. Decay theory of forgetting

Memories fade/erode with time
Brain cells die, connections weakened

2. Interference theory

New learning replaced or confused with old info
Proactive: caused by materials learnt prior to learning episode
Retroactive: caused by materials learnt after learning episode

3. Retrieval failure

Memory in long-term storage, but not located
May be partial, e.g. tip of tongue (TOT)

Forgetting: used to know (vs never acquiring)
Most memories accurate
Ebbinghaus forgetting curve: forget more over time (steep decline at start)

Undoing forgetting

Can 'undo' with retrieval cues & revisiting / context reinstatement - e.g. cognitive interviews
Not with hypnosis (more detailed, not accurate, reports)

Autobiographical memory

Episodes / events in own life, shape ID & behaviour
Same principles, but with self-reference effect + emotion + delay

Self-reference effect

More likely to remember info that refers to self
Because of deep processing, extra brain activity
Self-schemas
Bias how you remember past - make it more like present, match self-schema

Emotion & consolidation

Emotion important for consolidation (biological process to cement memory)
Emotion narrows focus
Traumatic: typically remembered well, or too well, due to consolidation - unless sleep deprivation, head injury, substance abuse, stress

Flashbulb memories

Clear, detailed, long-lasting, but may be inaccurate
Social aspects important: +ve and -ve (co-witness contamination)

Repression

Repressed memories unlikely
Repressed: traumatic, stored by not easily recalled
More likely due to retrieval failure, false memories

Over time

Longer retention interval -> lower recall
Well-learned or revisited -> less time impact
e.g. photos in yearbook: 90% recognition with little drop-off after 14 years, 50% recall, with drop-off

Lifespan retrieval curve

Childhood amnesia: few memories birth to 2-4yo
Reminiscence bump: clear, detailed (not always accurate) 10-30yo
Period of recency: remember recent
Applies to lots of diff memories
Similar across cultures, but US recall more childhood (self-focus), Maori recall more childhood (oral tradition)