Memory systems Flashcards
Experimental evidence for STM vs LTM (4)
Atkinson and Shiffrin (1968) - serial position curve, primacy and recency effects.
Glanzer and Cunitz (1966) - 10s filled delay between presentation and recall abolished recency effect (hindered rehearsal), preserved primacy.
Glanzer 1972- primacy effect affected by age of participant, length of words, rate of presentation. Recency unaffected.
Craik 1970 - 10 lists of 15 words, then asked to recall words on any of these tests. Negative recency effect observed, bc the last words from each list did not go into LTM.
Neuropsychological evidence for STM vs LTM (4)
Scoville and Milner 1959 - HM had normal digit span but impaired LTM, MTL damage
Shallice and Warrington 1970 - KF had normal LTM but impaired digit span, parieto-occipital damage
Schacter and Wagner 1999 - Reviewed 9 fMRI studies, encoding into LTM involves MTL
Hanson, 2000 - encoding into STM involves inferior frontal and parietal areas (i.e. near Broca’s, link to phonological loop)
Experimental evidence against STM vs LTM (5)
Bjork and Witten 1974 - 12s filled delay between every presentation preserves recency effect, so it’s bc recall is a backwards looking process
Baddeley and Hitch 1977 - Performing a digit span task during recall of a list of unrelated words preserves recency effect, so it’s not due to limited STM capacity
Baddeley and Hitch 1977 - Asked rugby players to remember teams they’d last played against, long-term recency effect unaffected by number of matches played
Pinto and Baddeley 1991 - Asked visitors to the APU to remember parking spaces from last 20 visits, long-term recency effect unaffected by how long ago these 20 visits were
Crowder 1976 - telephone pole analogy - as memory items get older they become less easily distinguishable. To account for primacy effect, he’s assuming that between each presentation you rehearse from the beginning of the list. Is this true?
Neuropsychological evidence against STM vs LTM (2)
Ranganath and D’Esposito 2001 - ask people to actively remember a new face A for 7s, then say whether face B was the same or different. Activity in MTL, so STM and LTM.
Ranganath and Blumfeld 2005 - studies of brain damage patients used stimuli too familiar (in studies using novel shapes and words, MTL damaged showed impairment even after 0s delay, and patients supposed to have only STM deficits found to extend to LTM). Maybe familiar stimuli are represented elsewhere in the brain.
Episodic buffer - necessity (3)
We need some store that can integrate different aspects of ST and LTM, because:
Why do we have a digit span of 7, but can remember meaningful sentences of 15 words? Original theory, use LTM info for meaningful ‘chunking’, but even densely amnesic patients can do it (Tulving, anecdotally - a patient with no LTM at all could remember all the cards in hands around a table of bridge)
Baddeley et al 1984 - Articulatory suppression (prevents visual stimuli from entering phonological loop) impairs recall of visually presented stimuli, but doesn’t abolish it as phonological loop theory would suggest (digit span drops from 7 to 5)
Patients with impaired short-term phonological memory could remember 1 aurally presented digit, or 4 visually. Could be visuospatial sketchpad - but this is thought to be better at a single complex pattern than serial presentations, and visual similarity effects were small and not limited to articulatory suppression (Logie et al 2000)
Chincotta et al, 1999 - Spatial task abolishes numeral over digit word advantage (bc numerals are VSS, words are PL)
Episodic buffer - what is it?
Limited capacity, probably distributed, accessed by consciousness, combines visuospatial and auditory information to form a long-term memory trace, holds LTM-retrieved info temporarily online in working memory, added to the Baddeley and Hitch Working Memory Model (1974) in 2000
Baddeley thought it may be a way to link WM to LTM. Phenomenological evidence suggests conscious monitoring of apparent memory evidence is crucial to avoid false memories and confabulation - maybe the episodic buffer does this?
Episodic buffer - experimental evidence (1)
Lehnert and Zimner (2006) - location memory similar for pure- and mixed-modality lists
de Pontes Nobre et al, 2013 - review of many studies into uni- and cross-modal WM. Problems: concurrent tasks designed to test attentional demands on EB may impair retrieval rather than encoding in the EB (because saying if two things were linked requires more attention than simple feature recognition). Also the concurrent task should be in a different modality to avoid interference.
Episodic buffer - neuropsychological evidence (2)
Prabhakaran et al 2000 - activity in right frontal lobe when individuals were asked to remember multi-modal, combined info. Activity more posteriorly when subjects were remembering unimodal info.
Simons et al 2005 - activity in similar frontal region during monitoring of info retrieved from LTM
Phonological loop - evidence
Phonological similarity effect - words that sound similar are harder to remember. Not true for visually or semantically similar words.
Word length effect - longer words are less well remembered, because they take longer to rehearse
Articulatory suppression - repeating an irrelevant sound like ‘the’ impairs performance. Also abolishes word length effect, because you can’t rehearse at all.
Transfer of info - adults, when told to remember visually presented stimuli, tend to say or mouth them, to transfer info to the phonological store.
Neuropsychological evidence - aphasic patients are impaired, because they can’t form motor codes for verbal rehearsal. Dysarthric patients (peripheral problem) are not impaired.
Experimental evidence for explicit vs implicit
Jacoby and Dallas 1981 - Deeper processing improves explicit memory but not implicit memory. Changing modality between learning and testing impairs implicit memory but not explicit memory.
What is explicit memory? How do we test it?
Memory that can be consciously accessed, tested using recognition, cued recall, free recall. Split into episodic and semantic.
What is implicit memory? How do we test it?
Memory revealed when performance on a task is facilitated in absence of conscious recollection. Tested using word stem completion, word fragment completion, degraded word naming. Split into perceptual representation system (inc priming, conditioning) and procedural memory (skill learning)
Neuropsychological evidence for explicit vs implicit
MTL amnesics like HM have impaired explicit but not implicit memory (e.g. HM learnt mirror drawing, but could not remember the training sessions happening)
Gabrieli et al 1995 - MS had occipital damage, and had impaired implicit but not explicit memory
Schacter and Wagner 1999 - reviewed 9 fMRI studies looking at encoding into and retrieval from explicit LTM, found activity in MTL
Simons et al 2003 - perceptual priming involved activity in fusiform cortex
Neuropsychological evidence for episodic vs semantic, BUT masked by widespread MTL damage
Vargha-Khadem 1997 - children with selective hippocampal damage, acquired at birth. Impaired episodic memory, but perfect semantic memory (so performed well at school).
Patient SS had extensive MTL damage and was impaired at all anterograde explicit memory. Patient PS had selective hippocampal damage and could acquire new semantic knowledge
Butters and Cermak 1986 - Patient PZ wrote an autobiography right before his lesion. Showed a temporal gradient in both in retrograde (events from autobiography, and scientific terms).
SPI model - what is it? In support of
Tulving 1998 - serial processing, parallel storage, independent retrieval
Proposed because of patients with impaired episodic but intact semantic, and contemporary lack of vice versa.
Monohierarchical - a memory is encoded perceptually, then semantically (for meaning and concept), then episodically (for involvement of self and time).
Anything that is to be stored in episodic memory must first be processed semantically - therefore there should be no patients with impaired semantic memory but intact episodic memory
A memory trace is a ‘bundle’ of widely but systematically stored traces, each bundle stored separately from the others, so can be retrieved independently
The model allows and even predicts a double dissociation between semantic and episodic memory in /retrograde/ amnesia
Allows lower-level memory to function in the absence of higher-level, e.g. babies can learn semantic knowledge without remembering when they learnt it, animals without higher-level brain development can learn things.
