Working Memory Flashcards
How is short term memory traditionally measured?
• Memory-span procedures:
• participant presented a sequence of items, required to repeat them back; start with 1 item, increasing number of items by 1 until participant make mistakes
• point at which the participant is able to recall all items correctly 50% of the time is designated as her/his memory span
Reading span task: originally developed by Daneman and Carpenter (1980). Task requires participants to read a series of sentences aloud and recall the final words
What are the various views on short term working memory?
- John Locke (1690) explicitly distinguished between temporary workspace for the “idea in view” and more permanent “storehouse of ideas”
- William James (1905) coined the expression “primary memory” as distinct from the storehouse of “secondary memory“
- Atkinson & Shiffrin (1968) referred to short-term memory as combination of storage and control processes
- Craik & Lockhart (1972) emphasised processing rather than structure in memory; nature or level of initial processing determines recall. Thus, memory was seen as a byproduct of cognitive processing rather than as a separate entity.
- working memory as multiple components: Baddeley (1992) refers to a brain system that provides temporary storage and manipulation of the information necessary for such complex cognitive tasks as language comprehension, learning and reasoning
What is the ‘Brown-Peterson” paradigm of short term memory (1959)
• participants had to recall trigrams (strings of three letters) at intervals of 3, 6, 9, 12, 15, or 18 seconds after the presentation of the last letter
after oral presentation of each trigram, participants asked to count backward by threes from a three-digit number spoken immediately after the trigram
What factors affect memory span?
• auditory presentation leads to larger memory span estimates than visual presentation
• rhythmic presentation is better than non-rhythmic presentation
recoding or chunking information; George Miller showed in his classic paper (1956) that memory span is determined by the number of ‘chunks’ or integrated items you need to recall, not the number of items presented
What is the model memory model?
Atkinson and Shiffrin (1968)
1. Short-Term Store Information maintained in an active state Verbal rehearsal (capacity = 7 ± 2)** Phonemic encoding 2. Long-Term Store More permanent record of experience Semantic encoding Cue-dependent Information is transferred from STS to LTS**
Goes through organs, once attended to goes into short term store.
Critical aspect of this model is the implication that the short term store is an anti chamber for memory proper. Must go through short term store in order to go into the long term model. Can only get into long term is stays in short term long enough and is rehearsed - this is critical aspect of their model.
Sensory stores decay rapidly. Short term store is forgetting via displacement. Capacity assumed to be very limited but unlimited in long term store and eventually becomes permanent.
What is evidence of the short term memory model?
• coding differences in long-term & short-term storage (Baddeley, 1964)
• brain-damaged (amnesic) patients:
1. had severe difficulty in learning new information but could recall information that they had learned prior to their injury
2. short-term buffer appeared to be intact but their long-term storage was grossly impaired (Teuber et al., 1968)
• different components of the serial-position curve in verbal free recall:
1. Immediate vs. Delayed Recall (decreased recency effect)
2. Speeded Lists (decreased primacy effect)
3. Incidental Learning (decreased primacy effect)
Had 3 lines of evidence supporting their models. Long term memory semantic coding.
Neuropsychology of brain damaged patients both ways. One incredibly short memory span (1 word eg) but long term memory still accessible as can remember things before the accident for example.
Others could never get things into long term memory - normal digit span in short term memory while rehearsing, but cant keep it once stop.
Their short-term buffer also appeared to be unimpaired in that they could store sequences of digits and maintain the sequence by verbal rehearsal. However, as soon as rehearsal stopped or was prevented, the sequence was forgotten. Short-term buffer also appeared to be unimpaired in that they could store sequences of digits and maintain the sequence by verbal rehearsal
What is recall performance?
Postman and Phillip (1965)
- Immediate vs. Delayed Recall
No distractor task after last item à Strong recency effect
15-s distractor task after last item à No recency effect - Final Free Recall
Recall all words from all lists
Plot # recalled from each serial position
“Negative recency” effect
Found in general task with immediate recall people good at first and last few words. First as rehearsed long enough having commanded attention from the start, and last few because recency effect still in short term store haven’t been displaced yet. If delay the recall, give distractive task for 30 seconds, get rid of their recently effect of those in short term store but still got those in the long term memory.
What were the three outcomes of free recall performance (Craik 1970)?
- Overt Rehearsal
Subjects asked to rehearse out loud
First few items receive the most rehearsal - Incidental Learning
Subjects unaware of impending memory test
No primacy effect - Speeded Lists
Less opportunity to rehearse
Primacy reduced, recency unaffected
What are the failures of the modal model?
• under this model, multiple working memory tasks should limit working memory capacity
• not the case; tasks having to do with lexical information do not limit the capacity to perform visualization tasks
• numerous other studies have shown that normal adult subjects can simultaneously carry out two tasks, each of which would apparently place heavy demands on a single flexible working memory system
• several studies have shown that estimates of the difficulty of performing tasks are on their own poor predictors of how readily those tasks can be performed concurrently
• K.F.: Suffered injury to the left parieto-occipital regions and showed severe limitations in verbal STM (digit span of 2)
• but K.F. could get information into LTM and retrieve info. How?
• information had to be presented visually; auditory information was not encoded
• thus, likely there are must be “multiple-components” to WM
• K.F. had an alternative route into long-term memory
(a) that did not involve working memory and/or
(b) there are several working memory subsystems, not all of which are damaged in patients such as K.F.
Too much focus on cognitive structures and not enough on processing information. People get good at certain types of tasks. Impossible with one short term store - either got to capacity or didn’t, cant use different modalities according to this model.
Can have incredibly localised lesions - eg oliver sachs
Seemed that there must be multiple components to the working memory and the model was far too simple.
Second, I shall argue that working memory is better thought of as a system that operates after access to long-term memory has taken place, rather than acting as a means of transport for sensory input to long-term memory. On this view, working memory is seen as a workspace rather than a gateway, and sensory input reaches working memory via long-term memory, not the other way around.
Suffered injury to the left parieto-occipital regions and showed severe limitations in verbal STM (digit span of 2)
K.F. could get information into LTM. How?
Information had to be presented visually Auditory information was not encoded Thus, there must be “multiple-components” to working memory
Despite having a digit span of just two items, K. F.
appeared to have normal long-term learning and retrieval.
How can we assess working memory?
Dual-task methodology
Participants perform 2 tasks at once with two assumptions
• if tasks use same component, they cannot be performed successfully together
• if two tasks use different components, it should be possible to perform them as well together as separately
Common results
• visual tasks interfere with visual retention (eg, football/driving)
• verbal tasks interfere with verbal retention (eg, articulatory suppression)
• demanding verbal/visual tasks interfere with reasoning (tap executive function)
Implications:
• multiple short-term memory systems
• articulatory loop vs. visuo-spatial scratchpad (VSSP)
About Baddeley’s working memory model
The original model of Baddeley & Hitch was composed of three main components; the central executive which acts as supervisory system and controls the flow of information from and to its slave systems: the phonological loop and the visuo-spatial sketchpad. The phonological loop stores verbal content, whereas the visuo-spatial sketchpad caters to visuo-spatial data. Both the slave systems only function as short-term storage centers. In 2000 Baddeley added a third slave system to his model, the episodic buffer.
What is the phonological loop?
- responsible for speech coding
- consists of two components: Phonological store and articulatory control process
Idea that the distinction between speech inputs and non speech inputs. Serial recall of verbal information. Get preferential putting away to short term store. This is about a 2 second store of memory and only way can keep it there is by subvocally rehearsing the information.
Assumed somehow important in language learning particularly developmentally. Sequential information is good - not good for seeing array of objects, just visual or speech input
How does the phonological loop work?
- lexical information is coded and stored in the phonological store for about 2 seconds
- the stored information is refreshed by means of sub-vocal rehearsal
- the sub-vocal rehearsal is the articulatory loop
- assumed to have developed on the basis of processes initially evolved for speech perception and production
- suited to retention of sequential information (memory span)
Determined jointly by durability of memory trace and time required to refresh trace by subvocal rehearsal
What are the limits of the phonological loop?
- capacity (items): Seven plus or minus 2?
- not necessarily
- Miller’s magic number does seem to hold (even with chunking), however…
- rehearsal rate is the main determinant of capacity
- capacity (time): about 2 seconds
Some people are better at chunking information. Large literature, is it decay or interference that limit ability to short term shore. Rehearsal rates is what keeps it going, quicker can get through the list of words, the better you can recall. If the words all sound the same people don’t tend to recall them as easily as get mixed up and the phonemics are too similar - letters or words. Hard words take time - both to articulate and store.
Preventing people from rehearsing, eg saying ‘the’ over and over again while visually presenting the words. Stops linking as cant rehearse any more and this disrupts the memory of works enormously.
What is evidence for the phonological loop?
- the Phonological Similarity Effect (GPVT harder than RHXK)
- the Word-Length Effect (More monosyllabic than polysyllable words remembered)
- Articulatory Suppression (Overt or covert inner speech disrupts lexical memory)
- the Unattended Speech Effect (Lexical-based sounds disrupts lexical based tasks)
Speech inputs preferentially access the store and can disrupt the memory
