Memory- using it not losing it Flashcards
Godden & Baddeley (1975)
Free recall of a list of 40 words was better when the environment (context) matched, whether underwater or on land
Morris et al. (1977)
- how were words encoded?
- When asked if items were old or new, which did people remember best?
- When cued with rhyming items, what did people remember better?
- Words were encoded using ‘deep’ vs ‘shallow’ study tasks
- When asked if items were old or new, people remembered semantically encoded ones best (generated to fit sentences)
- When cued with rhyming items, people remembered rhyme encoded ones best (generated to make rhymes)
What is an episodic memory made of?
- Details about an event eg. who was there
- Contextual information: time + location, what we were thinking
- Relations (associations) of details: people + time + location
- Reinstating part of a memory can help bring back the rest = contextual cueing / contextual reinstatement
Recall versus recognition tests
Free recall (minimal cue): List all your lecturers this year
Cued recall (cue is more informative): What courses does she teach on?
Recognition: Have you seen her before?
What can be a powerful contextual memory cue?
Location
Smith (2014)
Smith & Manzano (2010)
Study and results
Study: videos with sets of words
Test: recall by writing down as many words as possible
Test with video as cue, or not
Written free recall of words improved by reinstating images from scene videos at test
Memory was dramatically better with this cue (scene reinstated)
How do cues work?
- what type of memory?
- global matching models?
- complementary learning systems model:
- Content addressable memory – find by knowing content. Contrast with address addressable, e.g. where people live
- Global matching models: retrieval reflects the match between a cue and all stored memory traces (Clark & Gronlund, 1996)
- Complementary learning systems model: episodic memory representations stored in cortex, partial cue triggers pattern completion by the hippocampus (McLelland et al., 1995)
How do cues work?
- features of event, hippocampus combines them
- at point of retrieval get partial cue
- Then get whole process by pattern completion
- Reinstates memory representations that were stored
The pattern completion and pattern separation processes in the context of Teyler and DiScenna’s (1986) indexing theory. Pattern completion: (A) A set of neocortical patterns activated by a particular experience projects to the hippocampal formation and activates a unique set of synapses. The memory for the experiences is stored as strengthened connections among those hippocampal synapses activated by the input pattern (this is the index). (B) A subset of the initial input pattern can activate the index. (C) When this occurs, output from the hippocampal formation projects back to the neocortex to activate the entire pattern. Pattern separation: The hippocampal formation supports pattern separation by creating separate indices to similar input patterns. Note that two similar input patterns (ABCD and CDEF) converge on different representational units in thelowerlevel that represents the hippocampal formation. In contrast, these two patterns would not be separated in the neocortex, so it would have trouble keeping these patterns separated.
Encoding and retrieval:
1- how is context incorporated?
2- what helps retrieve memory?
3- what should a cue match?
4- encoding and retrieval are _____?
1- Context is incorporated in the memory trace
2- Cueing with context helps retrieve that memory
3- A cue should match – its processing overlap with – what was encoded (E.g. if you know your memory cue will be a photo of a location, you might encode that context better)
4- Encoding and retrieval are interdependent!- will make it easier to receive given that kind of cue
What did Tulving (1974) and Wiseman & Tulving (1976) show?
Tulving (1974): Cue-dependent Forgetting
Apparent ‘Retroactive Interference’ of new word list learning on Free Recall for old lists
But NOT seen in Cued Recall, when specific cues were given for each list
Wiseman & Tulving (1976): Encoding Specificity Principle
Retrieval depends on ‘Informational Overlap’ between stored information and information presented at time of retrieval
Smith & Manzano (2010)
Results with number of words tested
Study: videos with sets of words
Test: recall by writing down as many words as possible
Test with video as cue, or not
Results:
Scene cues more effective when each video context was studied with fewer words
Cues better when more diagnostic
Effectiveness of the cue depends on how unique or diagnostic the context is
What did Celia Harris argue about memory in lockdown?
When events in our lives become similar to each other as during lockdown, it’s more difficult to find distinctive (diagnostic) cues to retrieve individual event
Episodic reinstatement:
1- how are memory traces stored?
2- what happens if a partial cue overlaps with a memory trace?
3- what can this be shown by?
1- Memory traces are stored using some of the same neural representations that allow us to experience the events
2- If a partial cue overlaps with a memory trace it triggers recollection, and this reinstates the rest of the memory trace
3- This reinstatement can be shown with fMRI by looking for reactivation of neural patterns of experimental ‘events’
Episodic reinstatement:
1- what did Polyn et al. (2005) measure?
2- how were neural patterns during study discriminated?
3- what happened during recall?
4- how do we ‘relive the past’ during recollection
1- fMRI brain activity patterns when people studied and recalled faces, locations and objects
2- Machine-learning algorithms were trained to discriminate neural patterns during study
3- during recall the same algorithms could ‘read out’ (decode) what (which category) people were recalling
4- Neural reinstatement of memory contents
This is a type of multivariate pattern analysis
Remember that recall = the task, recollection = the episodic memory experience
Episodic encoding
1- what do events engage?
2- what is the PFC involved in
3- what are memories encoded as?
4- what does the hippocampus bind?
1- Events engage multiple areas of the cortex
2- organise information
3- a ‘byproduct’ of event processing
4- multi-element memory traces
Episodic retrieval:
1- what is episodic retrieval triggered by?
2- what does the hippocampus initiate?
3- what is reinstated?
4- what is the PFC involved in?
1- a cue
2- recollection in response to the cue
3- (some of) the original cortical activity is reinstated
4- organise & monitor
Polyn et al. (2005) show about brain activity patterns during recall
Reinstatement started about 5 sec before recall. Preliminary evidence that mental reinstatement (‘self-cueing’) actually triggers recall (but fMRI has poor time-resolution)
Shows that when people are recalling faces, the brain patterns are more similar to the original
Also, before people recall thew words, brain patterns are reinstated
This represented this process of self-cueing (mental representation)
Episodic reinstatement:
1- Different types of events have ______?
2- What happens when a cue triggers pattern completion?
3- What can a memory cue be?
4- application?
1- unique patterns of brain activity which get stored in memory traces
2- When a cue triggers pattern completion, reinstatement of these patterns is thought to support recall
3- external (eg a smell, a photo) or internal - deliberate mental reinstatement may also trigger pattern completion
4- Application: the Cognitive Interview for eyewitnesses
Diary studies: Linton (1978)
2-3 events recorded each day over 6 years
2 ‘semi-randomly’ picked later, monthly, for recall of details, dates. Delete event file if forgotten.
Qualitative study
When testing herself, 1 memory compared to 2 ect. after the same delay, items that had been repeatedly tested were better remembered
For 4+ tests, they remembered more than 70%
Galton (1883)
Breakfast table questionnaire – ask to recall a particular personal episode each day
What helps to remember things in the long term?
Testing your memory
The testing effect
Hanawalt (1937) results
Testing boosts memory at 1 week
SSSS = study 4 times for 5 mins
SSST = study then recall test
STTT = repeated recall tests
That’s 52% forgetting in the repeated study condition
versus 10% in the testing condition (STTT)!
On immediate memory tests studying can be better
(1) Semantic elaboration
- what enriches semantic representations of a memory, why?
- study? Carpenter (2009)
- what does this overlap with?
- Testing may enrich semantic representations of a memory because when we learn something, additional associations are formed around it, giving alternative retrieval routes
- study MOTHER–CHILD
- initial test with cue MOTHER: answer is CHILD
- later, can use associates of the cue, MOTHER, to retrieve the target
- E.g. FATHER: answer is CHILD – this answer is given more often after the practice test compared to a restudy condition
*this overlaps with the second account- by having these additional semantic association it gives you additional retrieval so more different cues
In the brain:
- Wing et al. (2013)
- what regions?
- BUT…
- encoding activity?
- Wing et al. (2013) memory encoding during study vs. initial test
- Hippocampus and temporal ‘semantic’ regions - only during testing
- BUT don’t know if semantic processing actually happening AND overall activity in these areas greater during restudy
- Successful encoding activity – showed that people were encoding during the initial test or restudy
(2) Episodic context
- when is the context different
- what updates context representations
- what now triggers recall?
- supported by?
- why?
- When item is studied and later tested, the context is different – not only place but also time, internal state
- Testing updates context representations so memory trace now includes both old and new context
- Larger range of potential cues can now trigger recall, as they may overlap with either old or new context
- Supported by findings that a difficult initial test is better e.g., recall MOTHER–C—- versus MOTHER—CHI–
- this is because people have to do more mental reinstatement (thinking back to place, time, internal state)
Explain Wimber et al. 2015 study + results
- Learn 2 word-face and word-object pairs
- Test with cues: antique–?
- Practice tests to improve memory for only first pairs
- Indirect fMRI support for context account, strengthen relevant & weaken irrelevant object representations
- Stronger activation for tested items
- Suppression for competitor items
- (In cortical regions that represent face/object memories)
Testing strengthens relevant representations and suppresses irrelevant representations, so reducing the number of associations that retrieval cues trigger - dotted lines are best linear fit
The more they practiced trying to remember which face went with antique, there was a gradual reduction in representations
Explain Sekeres et al. (2016) study
Memory for video clips tested over 7 days
Given 40 film clips, then another 40
Higher forgetting of peripheral (perceptual) details than central (gist) information
But cues (reminders = R) re-triggered recollection of the peripheral information
Repeated testing also reduced forgetting of peripheral details
Found that generally, people remembered the gist and forgot the detail
- what does the testing effect show?
- what does this suggest about the variables that assist initial encoding
- The testing effect shows that a memory is updated in some way when it is retrieved
- Therefore, the variables that assist initial encoding may NOT be the same ones that benefit longer-term learning.
Richards & Frankland (2017)
(maps)
What did they find?
Getting people to learn where certain places where
Store a memory: Kyles bistro is S-W from home
Learn something new: Kyle’s Bistro has moved to N-E
Found- forgot old memory to promote flexibility
Make a prediction: new restaurants often appear to the south
Found- forgot particulars to generalise
To make inference, you need an updated memory
Systems consolidation vs Multiple memory traces
Systems consolidation – memories become independent of the hippocampus with time
Multiple memory traces – updating over time. Hippocampus always involved
In either case generalised, context-independent (semantic) memories develop over time
Some episodic memories may become semanticised
Systems consolidation doesn’t differentiate semantic vs episodic and consolidation may or may not involve ‘semanticisation’
