Learning, Memory And Forgetting Flashcards
What is memory?
Common view is that its a ‘store’ from which we retrieve facts
What does memory depend on?
How do you remember something in the future, as it hasn’t happened yet, so isn’t in the store
How do people come to misremember things that never happened - aren’t in the store but people claim to remember them
What’s an alternative definition of memory?
Natural inference system that allows us to store a few facts and derive others as needed.
Relate new events to prior knowledge in order to understand them.
Deliver relevant knowledge when it is needed.
What are the types of memory?
Episodic memory:
- specific details of events/episodes e.g. ‘I remember when I had breakfast this morning’
Semantic memory:
- our ability to store facts and categories e.g.‘ I know that a canary is a bird …’
Procedural memory:
- memory regarding skills, e.g., learning to walk, learning to type, riding a bicycle (also one form of non-declarative memory)
What are basic memory processes?
Encoding
- Code information and put into memory (acoustic codes and visual codes)
Storage
- Maintenance of information in memory
Retrieval
- Recovering information from memory and bringing it into consciousness
Whats the multi-store model of memory?
Atkinson and shiffrin developed a model involving 3 stages of processing
Memory stores form the basic structure.
Processes such as attention and rehearsal control the flow of information between them. BUT main emphasis was on structure.
What stores are in the multi-store memory model?
Sensory
Short term
Long term
Whats the sensory store?
Both visual and auditory stores “hold” more information than we can process
We need to attend to certain items in order to remember them
Whats the short-term memory store?
Limited capacity:
- Evidence: Measures of digit span (how many numbers people can recall).
Fragility of storage:
- Evidence: Information decays rapidly from the short-term store, with information usually decaying within 18 seconds (as illustrated in the Brown-Peterson Paradigm, which we’ll discuss shortly).
- Evidence: Interference effects in the short-term store (as illustrated in the release from proactive interference, which we’ll also discuss).
Whats limited capacity - digit span?
According to Miller (1956), people can remember 7 +/- 2 chunks of information.
The word chunk is crucial—you can improve your performance on digit span tasks by grouping digits together into chunks.
For instance, Ericcsson et al. (1980) reported a study in which a participant improved his memory span to 79 digits, by chunking the digits together.
The usefulness of chunking information as a recall aid lies behind the way phone numbers are organized—e.g., my office number 9284 6352.
There is still ongoing debate over how many chunks we can recall (Cowan, 2005).
Whats the fragility of storage - decay?
Evidence: Brown-Peterson Paradigm (named after the authors Brown, 1958; Peterson & Peterson, 1959)
- Participants asked to study triads consisting
of sets of 3 letters (e.g. TVG, KLM). - Participants counted backward in threes for a
short period (to prevent rehearsal of the
triads). - Participants tried to recall the original triads.
- Performance declined with the delay
before recall.
Whats the fragility of storage - interference effects?
However, simple decay isn’t the only means through which information becomes displaced from the short-term store.
There are also interference effects, meaning that other information interferes with recalling the information you are trying to learn or recall.
Interference can take the form of either proactive interference (competition from previously learned information) or retroactive interference (competition from subsequently learned information).
What’s proactive interference?
Proactive interference means that people have trouble learning new material because previously learned material interferes.
If you are asked to recall the triads of XCJ, HBR, KSY, then it would be difficult to recall a 4th triad of KRN, because the older items interfere with your ability to learn new items.
BUT if the last item is 164, then the task is easier because the triads don’t ‘interfere’ remembering a three-digit number in the same way they interfere with remembering another triad.
Evidence: Wickens et al. (1963) - the release from proactive interference.
Whats release from Proactive interference?
Task = to remember
“XCJ, HBR, KSY”
The decline in performance because items are similar and interfering with each other in short-term store.
But when last item is different (e.g. “164”) then interference disappears.
But when final item is “KRN” then performance continues to decline.
This is known as “Release from proactive interference”
What form is information encoded in the short term store?
Early view was that information in short-term memory was held in an acoustic, verbal code (i.e. speech-like).
Subsequent research showed that, in fact, the short-term store could retain semantic (meaning-based) and visual codes too.
Evidence for acoustic and verbal encoding?
Evidence: Wickelgren (1965) had participants study a list of letters. They then copied down another list of letters, which either sounded similar to the letters on the original list or did not sound similar to them. Finally, participants were asked to recall the initial list of letters.
Recall for the initial list of letters was worse if the intervening list of letters sounded similar, indicating that this acoustic similarity created greater retroactive interference.
Evidence for semantic encoding?
Evidence: Wickens (1972)
Release from proactive interference when different semantic categories were used
e.g. all pts asked to remember different kinds of fruits, then performance declines due to interference
Except when categories switch on the last trial - then release from PI (except for pts still asked to remember fruits)
Whats working memory?
The information processing approach sees STM as a ‘short store’ of information that is later transported into LTM.
Baddeley & Hitch (1974; see also Baddeley, 2012, for an updated account) proposed a different view of STM which focuses on how information is used, not just stored.
But this idea did not just spring into life from nowhere - what was the evidence that B+H based their conclusions on?
Based on inadequacies with the simple “store” model of memory (specifically the idea of 7 +/- 2 “slots”).
Whats the problem with storage-only model?
By the mid-1970s all sorts of role and functions were being attributed to the short-term store, including problem solving, comprehension and reasoning.
But could a short-term store capable of storing 7 +/- 2 items of information really keep track of such complex tasks?
Can a simple 7 +/- 2 system cope with all the keeping-track and problem-solving processes here?
Whats early evidence that the 7 +/- 2 store might not be sufficient?
Evidence: Case studies of individuals with specific impairments in different types of short term memory recall (Baddeley, Papagno & Vallar, 1988).
An individual (PV) who had a digit span of just two items which would suggest that her entire short term memory system was severely impaired / damaged?
If this is the case then her other memory, comprehension, reasoning and problem-solving skills should be similarly impaired.
But they were not!
- This means that short-term memory must consist of something other than just a “store” - there must be other components capable of performing these other tasks.
What is the working memory comprised of?
Articulatory rehearsal loop / Phonological loop:
- deals with the rehearsal and retention of verbal phonetic information (like the loop of an audio tape).
Visuo-spatial sketchpad:
- deals with the retention of visual information.
Executive control system / Central executive:
- supervises and regulates information in working memory. Allocates mental resources to tasks.
What are the predictions of working memory model?
If 2 tasks make use of the same component (say, the phonological loop), they cannot be successfully performed together (they would interfere with each other).
If 2 tasks make use of different components (the first one uses the phonological loop and the second one uses the visuospatial sketchpad), it should be possible to perform them together relatively well, though still with some decline in performance (as we saw in the Attention lecture last week, there is decline in performance even when attending to two tasks in different modalities).
Examples of working memory model predictions?
Listening and talking make use of the same working memory component – thus there is almost complete competition for global mental resources which increases your cognitive load and therefore reduces performance drastically.
Listening and doodling make use of separate working memory components – thus there is less competition for global mental resources.
What evidence is there for three components of WMM?
Neurological evidence
Experimental evidence
Whats neurological evidence for 3 components for WMM
From people with very specific deficits in certain cognitive abilities (say, the ability to recall sequences of digits).
For example, patient PV who had a digit span of only two but was otherwise unimpaired (although couldn’t learn new languages; Baddeley et al., 1988).
This suggests that the part of working memory dealing with the recall of digits can be selectively impaired. If so, this suggests it is a separate component.
Whats experimental evidence for 3 components of WMM?
People with very specific cognitive impairments like PV are rare. Therefore Baddeley and Hitch decided to manufacture their own patients using student volunteers.
They did this, not by removing the relevant part of their brain, but by “functionally disabling it by requiring participants to do a concurrent task that was likely to occupy the limited-capacity short-term storage system to varying degrees.” (Baddeley, 2012, p. 6).
For example, by asking people to do a visually presented grammatical reasoning task while hearing and attempting to recall digit sequences of varying length.
They found that response time on the reasoning task slowed as the length of the digit sequence increased, but that people could still perform the task, if not as easily. This suggested that the reasoning task included …”a clear involvement of whatever system underpins digit span, but not a crucial one.” (Baddeley, 2012, p. 6).
What’s the phonological loop?
Phonological store
- holds acoustic or speech-based information for about 2 seconds.
Articulatory store
- Process which produces ‘inner speech’ we all hear (also sub vocal articulation)
Whats the evidence for phonological loop?
Word length effect (Baddeley, 1975):
– Subjects asked to recall sets of words in the correct order.
– Their ability to do so was better for short words than for long words
Whats the visuo-spatial sketchpad?
Responsible for the setting up and manipulating of visuospatial images and is separate from phonological loop.
Whats the evidence for visuo-spatial sketchpad?
Evidence: Quinn and McConnell (1996)
- Participants learned a list of words using either visual imagery (uses VSS-P?) or rote rehearsal (uses PL?)
– Task performed either own its own or in the presence of dynamic visual noise or irrelevant foreign language speech.
– Results supportive of a separate visuospatial system.
Evidence: (Quinn & McConnell, 1996)
- Graph shows that dynamic visual noise interferes when the task involves the visuospatial sketchpad but not the phonological loop.
- And vice versa, that hearing irrelevant speech interferes when the task uses the phonological loop, but does not interfere with tasks involving the visuospatial sketchpad.
What’s the central executive?
Initial specification quite vague…relating to attention (the supervisory attentional system, or SAS)
Baddeley (1996) suggests a more specific definition:
– Switching of retrieval plans.
– Timesharing of dual-task studies.
– Selective attention to certain stimuli while ignoring others.
– Temporary activation of long-term memory.
Whats the evidence for central executive?
Random number generation studies (e.g., Teasdale et al., 1995):
– Participants asked to generate one digit every second in a random sequence.
– Approximately every 2 minutes, the researcher interrupted the task and participants wrote down their thoughts (if any).
– On trials where participants had no situation independent thoughts (SITs), the numbers were really random.
– On trials where participants reported SITs, the number sequences were not truly random.
When there were no specific thoughts, this indicated that the CE was not active. Nothing organising thought - therefore numbers random.
When participants reported independent thoughts, this indicated the CE was active and organising things - therefore numbers were not truly random, but organised in some fashion.
What are forgetting rates?
Ebbinghaus (1885/1913) pioneered the systematic study of human memory.
He taught himself lists of “nonsense syllables” (like LEV and BUP).
He re-learned these list after varying time intervals, from several minutes up to one month later.
He used the amount of time required to re-learn each list as a measure of how much was forgotten
Ebbinghaus found that forgetting is initially rapid and then slows markedly - forgetting curve
Whats the interference theory?
Proactive interference
- The tendency for older memories to interfere with the retrieval of more recent experiences and knowledge.
- calling your current partner by your ex partner’s name
Retroactive interference
- Forgetting caused by encoding new traces into memory in between the initial encoding of the target and when it is tested
- First-language attrition (Isurin & McDonald, 2001)
Whats cue dependant forgetting: the encoding specificity principle?
Forgetting that occurs because we lack the appropriate cues
Tulving’s (1976) Encoding Specificity Principle:
- Retrieval success is directly related to the degree of informational overlap between the information presented at retrieval and the information stored in memory, including its context
Whats an illustration of the importance of cues?
Tulving and Pearlstone (1966) had participants learn words which were presented as being part of different word categories (e.g., animal words).
When later asked to recall the words, participants were either (1) cued by given the category names or (2) just asked to freely recall the words.
Recall was better when participants were given the category cue
