Karteikarten
DD303 Working Memory
Karten in diesem Stapel (24)
- 1Problems with modal models of memory?eg Atkinson & Shiffrin ('68)
>Assumption - to achieve L/T learning info has to pass thru STM
>Assumption - just having info in STM guarantees it moving to LTM. Evidence shows retention not related to how long info held in STM but what you do with it - eg levels of processing (Craik and Lockhart '72)
Neuropsychological:
>LTM amnesic patients (eg HM) don't acquire new info yet have normal STM
>STM amnesic patients (eg KF - Shallice & Warrington) poor digit span, normal LTM
>Impaired STM should impair LTM & vice-versa
Estimates of STM capacity vary w/ no explanation - 2Type of experimental method used by Baddeley & Hitch ('74)?To try and mimic deficits seen in neuropsychological cases
Dual task method - assume limited S/T store - more it's utilised, less resource available for other task
>If 2 tasks use same component, they cant be performed successfully together
>If 2 tasks use different components, it should be possible to perform them as well together as separately
x2 tasks verbal vs x1 task verbal & x1 task visuo-spatial
Ps did reasoning Qs whilst trying to hold up to 6 digit number in head
Longer time to answer Q with longer digit - 3Initial 3 components of Baddeley & Hitch WM model?Central Executive - controls:
Articulatory Loop
Visuo-spatial Sketchpad - 4Alternate WM model?Cowan ('88)
WM = activated component of LTM
WM performance affected by what we know
Agrees w/ B&H re central executive (ie SAS) - 5Example of 3 components working together (hint - windows)How many windows does your house have?
CE = strategy & running it
VSS = image of house
PL = counting windows - 6Phonemic similarity?Harder to remember list of words (in order) when they sound the same
Articulatory rehearsal loop - store 2-3 speech items
Central exec. - process info & store while processing
Processing & storage compete for limited resource
NB - if more words to learn and longer to learn them - meaning more important
Suggests PL interested in sound not meaning - 7Word length effect?Assume sub-vocal rehearsal
>Can remember more shorter words than longer ones
>Those who speak faster can remember more words
>Faster spoken languages remember more words - Naveh-Benjamin & Ayres ('86)
Rehearsal period = 2 seconds - 8Articulatory suppression?Part of dual-task method
Having P repeat "blah" whilst trying to remember word list
Stops rehearsal - Eliminates:
Word length effect
Phonemic similarity effect - 9Baddeley's personal experience of visuo-spatial sketchpad performance?Driving whilst listening to football game (in the US)
Gets engrossed, car starts weaving!
Switched off radio & okay
Tested in UK - effect of holding visual image is inconsistent w/ motor task - 10Developmental aspects of phonological loop?Young children = poorer recall spoken stimuli w/ long names than same stimuli w/ pictures - Hitch
Baddeley et al ('98) - Children's auditory span & vocabulary size correlated
Gathercole et al ('97) - PL capacity good predictor of vocabulary problems
BUT most studies correlational
Linked to learning to read - Dyslexics poor digit span & performance on PL tests - 11Theoretical issues with phonological working memory?Model doesn't explain irrelevant speech
Lovatt et al - word length effect not related to spoken duration. Can be eliminated if complexity controlled for
Hulme et al ('84) - Young children have word length effect but not yet learned rehearsal strategies
Cowan et al ('92) - word length effect caused by output delays (w/o rehearsal) - 12Key WM dates1968 - Atkinson & Shiffrin: modal model
1974 - Baddeley & Hitch: tripartite model
1984 - Baddeley: renames articulatory loop to phonological loop (phonological store & continuous sub-vocalisation process)
1986 - Baddeley - adopts Normal & Shallice concept of Supervisory Attentional System (SAS) for central executive
1996 - Baddeley - abandons storage in executive, fractionates it into modules for attention
2000 - Baddeley - multi-modal Episodic Buffer added - 13Evidence for fractionation of Central Executive (1996)Alzheimer's patients - concurrent tasks hard vs normal aging
Miiyake et al ('00)
3 executive functions/processes partially overlapping Baddeley's
1) Inhibition function
2) Shifting function
3) Updating function
Baddeley function - dual-task processing
More research req'd - 14Coherence & binding problem?WM model fractionated
How does info from sub-systems combine? (Jones '93)
Remembering words: 5-6 individual words; more if part of a sentence
Where are they held together & how?
Baddeley ('00)
>Multi-modal episodic buffer
>Combines info from sub-systems
>Temporary storage - allow sub-systems to interact with LTM & perception
>Output = episodes/chunks (3-4 chunk capacity)
>Seat of conscious awareness - 15Evolutionary purpose of Phonological Loop?Language acquisition
Neuropsych evidence:
Baddeley et al ('88) - Patient PV, poor digit span, normal intelligence & laguage
Not able to learn Russian/Italian word pairs
Suggest PL req'd for learning language
Experiment: Papagno ('92) - Adults learn word/non-word pairs; increase phonemic similarity or # syllables in non-words impairs learning
Word pairs = no effect
Suggest PL req'd to learn novel words
Individual differences: Service ('92) PL capacity correlates w/ vocab in 2nd language - 16Daneman & Carpenter ('84)>Individual differences in WM & language
>Measure storage combined w/ manipulation = WM span
>Read sentences & remember last word of each
>Complex: typical span = 3
Predicts:
>comprehension
>learning new computer language
>attention focus
>performance in std intelligence test - 17Logie ('95) - Visuo-spatial Sketchpad2 separate systems:
Visual Cache - info re visual form & colour
Inner Scribe - spatial & movement info
Beschin ('97) - patient NL's performance consistent w/ someone w/ selective damage to visual cache - 18Visuo-spatial sketchpad tests?Corsi Blocks test (Corsi '72) - P follows & copies sequence of blocks
Measure of visuo-spatial WM span; similar to digit span for PL WM capacity - 19Evidence for the visuo-spatial sketchpad?De Renzi & Nichelli ('75) - Corsi span & auditory digit span could be impaired independently in patients
Baddeley & Lieberman ('80) - visual mnemonics disrupted by spatial task NOT by visual task
Farah et al ('88) - Patient LH = spatial abilities better than visual
Williams Syndrome - impairment of visuo-spatial processing but normal verbal processing (Bellugi et al '94) - 20WM model - current status>One of most successful models in cognitive psychology. Widely cited each year
>Lot know about phonological loop/store
>Little known about visuo-spatial sketchpad & its role in movement & imagery
>CE still focus of work
>No attempt to explain automatic processing
>Link with LTM - other than an interface? - 21WM & garden path sentences?>Assumption - multiple interpretations of ambiguous sentences held in WM (Just and Carpenter, 1992)
>Just and Carpenter: individuals with low WM capacity less able to maintain multiple interpretations than individuals with high WM capacity
>Alternative theory - comprehension draws on more specialized resources than WM (e.g. Caplan and Waters, 1999). - 22Evidence suggesting the need for the Episodic Buffer?>Tulving - Bridge playing amnesic
>Prose recall - normal & amnesic Ps
>Resistance in serial recall (of visually presented numbers) to articulatory suppression - 23Consciousness & the Episodic Buffer?>LTM & Central Executive major areas
>WM = useful empirical tool w/ limitations
>Previous model - central exec key role in binding problem & integration BUT it doesn't have a store
>Previous model - no means of storing complex images > big role for LTM
>Suggests further store capable of getting info from slave stores & LTM & holding it in integrated form - 24Central Executive - central workspace?>Towse et al ('98)
>Test kids reading, operational & counting spans w/ parallel testing
>Hypothesis - as test progresses more storage req'd = less resource for processing
>Results - no decline in processing
>Alternative hypothesis - rather than share attention between processing & storage, attention switched to/from processing & storage
>Test - change time intervals for storing info
>Results - Spans lower as time interval increased
>Similar effects in adults when manipulate order of presentation (Towse '00)
>BUT Hitch el at ('01) weak tendency for processing to slow as storage load increases
SUGGESTS mixed model w/ attention switching & resource sharing
WM & central executive not simple construct - linked to fractionation of central exec.