Retrieval of Memory Flashcards
What does the DMN do when you are focused on a hard externally oriented task?
Deactivates relative to baseline
Why do source memory tasks activate ventrolateral PFC?
There is competition between sources -> judgement you’re making about the size and object are semantic
Which of the following does NOT cause strong interference between memories?
Competition memories which is highly meaningful -> as usually reduces the competition you have memories (basis of the level of processing effect -> think about the meaning of words, it tends to improve memory by reducing competition because it encodes items in a more distinctive fashion)
What type of interference occurs when new learning makes older memories inaccessible?
Retroactive
What circumstances promote false memories?
When there is competition e.g. from semantic knowledge
Why does damage to the hippocampus not promote false memories?
because if it’s already damaged retrieving memories will be hard anyways
What type of semantic impairment is associated with false memories?
Semantic Aphasia
What two deficits are associated with confabulation?
Poor control of memory and a meta-cognitive deficit
Why is retrieval induced forgetting helpful?
Reduces need to control retrieval of target item
Overtime, where may episodic experiences transfer to?
temporal neocortex
-> semantic hub in the anterior temporal lobe
How are memories retrieved?
often spontaneously
how is retrieval from strong cues?
automatic (due to spreading activation)
how is retrieval from weak cues?
can be effortful and we can struggle to separate memories
what was the DMN originally described as?
task negative network / believed to originally deactivate during a demanding task
what are the components of the default mode network?
incl hippocampus, angular gyrus, medial pfc and posterior cingulate
The DMN is implicated in?
spontaneous thought and episodic recollection (/recollective experience)
what happens to the DMN during a difficult task?
often deactivates to support memory retrieval
-> trying to do this during a difficult sensorimotor task, it will interfere with task performance so you might suppress this network deliberate during a non-memory based task
-> if you’re doing a memory task, this network may allow you to do that task
when are regions of DMN active more?
when focused memory on memory -> dependent on the task
what does DMN support?
deliberate and accidental memory retrieval
What did Christoff et al. (2009) find?
more DMN activity during a task where people reported their mind had wandered
Brain Map
Red regions are task positive regions -> show an increased response on a demanding sensory motor task -> this is also important for memory as tends to be switched on during encoding
- Red (encoding) -> Blue (Retrieval -> DMN)
what does episodic memory overlap more with?
episodic memory overlaps with DMN more than semantic memory
Why does episodic memory overlap with DMN more than semantic memory? (Vatansever et al. (2021))
- contrast people making episodic and semantic decisions about word pairs
- in episodic task -> if you contrast that with semantic task, you see more activation in areas of the DMN (posterior cingulate and angular gyrus)
whereas, if you everse contrast and you look at semantic decisions over episodic, you get a different set of areas including the ventrolateral PFC (important for memory control)
what is episodic memories about?
recently formed memories
what is semantic about?
making decisions about something you have gained knowledge about over a life-time ago
Vantansever et al. (2021) manipulated difficulty of episodic and retrieval tasks in similar ways (comparing strong and weak memories across tasks).
- made semantic tasks harder by asking people to make decisions about word pairs, with rarely co-occur (i.e. dog-bone (easy) vs. dog-beach (takes some thinking to see them together)
- episodic presented many learning tasks (lots of opportunities to learn the two words are going together in an episodic memory paradigm, then you make it easy to retrieve that information later) vs. making the decision more difficult by showing the individual a word pair only one
What did they find?
- same brain regions were important for easy decisions over hard
- DMN was implicated in both semantic and episodic retrieval when it was relative easy
- Angular gyrus and posterior cingulate might be important in relatively automatic patterns of retrieval from both memory stores
- WHEN WE LOOK AT HARD OVER EASY: we found VLPFC area (same region playing a role when you’re controlling retrieval from both semantic and episodic memory)
what is DMN involved in?
not only episodic recollection but also more automatic semantic retrieval
* DMN is recruited when memories are strong and come unbidden to mind (doesn’t take a lot of retrieval effort)
Where is the VLPFC located?
in the ventral and lateral part of the prefrontal cortex.
when is the ventrolateral PFC recruited?
during both episodic and semantic retrieval when task-relevant information is weaker
-> VLPFC when together with other regions in a large scale network is recruited when both episodic and semantic retrieval is more challenging when perhaps the information relevant to your goal is weakly encoded and therefore hard to recover
what issues do we have with retrieval?
we have many similar memories that interfere with each other
-> we must control competition between memories
what engages when we have to control competition between memories?
VLPFC
what is proactive interference?
old learning interferes with new learning
-> many similar memories generate strong competition
what is retroactive interference?
new learning interferes with old memory
what conditions maximise interference between memories?
- competing memory was presented recently
- competing memory is very similar to target
- there are many competitors
the VLPFC plays a critical role in what?
control of memory
but what is VLPFC also important for
selective coding and selective retrieval
what are other names for the ventrolateral PFC?
Left Inferior Frontal Gyrus (anatomical terms) or Broca’s area (tends to be in language studies)
what can the VLPFC biases?
processing towards the representation that is relevant for your goal/conext
-> mechanism allowing you to prioritise some memory representations over others as you go forward
how does encoding of memory work?
you have the capacity to use attention to guide the encoding of memory
-> but if you don’t pay attention, you won’t remember later
Subsequent Memory Effect
Research has looked at the basis of memory retrieval, using a paradigm called the subsequent memory effect
* Look at which brain regions during encoding predict subsequent memory success
* Scanning while they encode the information and then ask them the information they have encoded afterwards
* Can predict later success in memory task based on how much activation you see during encoding
Where does brain activity at encoding predict subsequent memory success?
VLPFC and Hippocampus (/Parahippocampal Gyri)
-> Hippocampus predicts subsequent memory encoding because the change things/generate activity in order to form an episodic memory that can be remembered later
-> PLIFG also predicts sequent memory encoding, important for predicting memory based on the way it can direct our attention to important aspects of our experience to allow selective coding of that information
-> VLPFC directs attention to important aspects of experience
* Attention to features that make it into your memory, as opposed to the memory system itself
if those with amnesia were doing the subsequent memory task, what would we find?
hippocampal response will be eliminated but VLPF response will still be there -> still engage attentional systems in memory in a relatively normal way even if they don’t really encode much information about what is going on
(Dobbins & Wagner, 2005) Classical Paradigm looking at interference in memory retrieval and the resolution of that inference by the VLPFC
- Participants saw pictures and had to make one of 2 decisions about the object: either the size, or the meaning of the object (living or non-living)
- Counterbalanced
- Later on, participants are asked to make decisions in the scanner
- In source memory condition, they were asked which coding task, they had previously seen the item (bigger-smaller task or the living-non-living task) -> lots of competition with encoding (environment)
- Can contrast this source memory decisions (which require resolution of competition) which novelty decisions (deciding whether they had seen the picture between - choosing the one which is unfamiliar)
What is seen if you contrast them?
VLPFC shows much stronger activation during source memory because of its role in resolving interference
-> VLPFC important for controlling competition inherent to source memory
How is the VLPFC crucial for controlling competition between memories?
VLPFC crucial for retrieving the link between objects (i.e. your bicycle) and specific context (where you left it this morning) when there are competing contents
while interference blocks access to memories, what can it also cause?
false memories
-> causes us to recall inaccurate information and semantic memory leads you to search in the wrong location
(Badre & Wagner) Neural Basis of Interference Resolution
- Paradigms show in inference situations from other similar experiences, you get more VLPFC activation again
- Participants saw 4 words and had to decide if a subsequent word was in the set or not
- Sometimes prevented items which were in the previous set i.e. peace in trial 2 (real hard to reject because it was in the previous set, causing competition from a recent experience -> this is now as a negative recent trial)
- BUT if the word (like trial 3) is not in the previous set, it’s a lot more novel and called a negative non-recent trial
What did they find if you contrast recent negative trials (hard) with non-recent trials (easy)
you get VLPFC activation again
What does the neural basis of interference resolution suggest for those that can’t activate the VLPFC in a normal fashion?
will experience more interference between memories leading to false memories (recollections of experiences which have never occurred)
what might happen to memories of those who have prefrontal cortex injuries?
have a lot more dramatic false memories
Retrieval failures following damage to PFC?
- memories aren’t ‘lost’ as in amnesia - they might not be readily retrieved -> retrieval may be disordered or impaired in some way
- poorer recall than recognition and stronger effects of cueing [cued memories reduces retrieval/control demands of a task]
- poor source memory - problems discriminating between similar memories
- unhelpful information is retrieved - false memories
false memories following damage to PFC: what is confabulation?
- failure to inhibit irrelevant thoughts/memories
- plus failure to detect implausible responses - this is a meta-cognitive deficit (which links to the medial PFC)
how do normal people evaluate false memories?
normal people: very vulnerable to false memories because of the way your expectation feeds into your episodic memory but we constantly evaluate how likely this is. If you can’t access specific sensnory details or verify multiple episodes, you’ll reject the idea as it’s not something worth pursuing
BUT maybe this doesn’t happen for those with damage, maybe they cannot evaluate how true their cognition is
what is Korsakoff’s syndrome?
Amnesia associated with long-term alcoholism
* caused by thiamine (vitamin B1) deficiency
* drinkers may neglect their diets
* alcohol interferes with conversation of thiamine into active form
what damage does Korsakoff cause?
wide spread to mammillary bodies are severely affected, plus prefrontal cortex
what are some behavioural symptoms of Korsakoff’s
- behavioural changes i.e. apathy
- problems with retrieval as well as formation of memory - retrieval can be cued
- confabulation
are recognition tasks always sufficient?
Sometimes recognition tasks are insufficient in recovering information in the case of hippocampal amnesia patients
what is wrong with retrieval in Korsakoff’s syndrome?
-> cannot retrieve information is a flexible and controlled way
* With a cue, more information can be retrieved/recall, evidence there’s a retrieval problem and a difficulty in maintaining memories themselves
* Cue should reduce retrieval demands -> someone has memories that are actually preserved but they couldn’t quite access
how is competition controlled?
by suppressing one memory in favour of another
-> VLPFC biases process towards features which are relevant
what does semantic memory involved?
many different stored features of our semantic knowledge (i.e. visual, sound, feels and associations) -> represented in the anterior temporal cortex
But this store of knowledge might not be sufficient for successful semantic cognition (using semantic information to generate appropriate thoughts and behaviours)
.e. want to take dog for a walk, this association that dogs need walk is relevant and fairly easy to retrieve as it’s strongly encoded. -> relatively automatic and the DMN might support this aspect of retrieval
BUT what if someone talks about dalmination at a Cinema -> you can come up with an association through control processes so you can recover relevant information for that particular context i.e. 101 dalminations film
like episodic memory, in order to recover the specific information you need for the content of a particular goal -> might have to apply control processes to suppress other competitive information
(Jefferies and Lambon Ralph, 2006) Contrasting different groups of patients who have damaged different brain regions -> to support the control of memory retrieval
* comparing patients with semantic dementia (anterior temporal atrophy) vs. stroke (ventral lateral prefrontal cortex)
* these two groups of people can have the same problems on the same task but for many different regions
But what is the difference on how they perform in these tasks?
patients fail the same range of verbal and non-verbal semantic tests but for different reasons
- Dementia: losing their knowledge
- Aphasia: can’t retrieve the relevant information even though the concepts are preserved
In comparison of these two groups of patients, what did they find in dementia patients when asked to name pictures?
Loss of the concept of low frequency animals (that they don’t talk about frequently)
* They maintain some superordinate understanding and names of frequently encountered concepts yet lose specific information about particular animals
[SD: progressive breakdown in knowledge from specific to general information]
In comparison of these two groups of patients, what did they find in semantic aphasia patients when asked to name pictures?
- Produce somatic or associated errors (i.e. zoo) -> requires a lot of knowledge to say zoo
- Also produce responses at the wrong level (i.e. Nelly the Elephant)
- Produce a semantic coordinate i.e. Rhino -> this would not happen with dementia because they would not know this concept exists and would be so semantically degraded that there would be no competition
[retrieval dominated by strong associations, even when these are irrelevant]
what is the difference between semantic dementia and aphasia in picture naming errors?
There is a breakdown of knowledge in semantic dementia. Whereas, in semantic aphasia, the retrieval is dominated by the most strong associations/similar items in memory even when they are not relevant
what may help individuals in picture naming tasks?
Cueing can help (phonological -> sounding words bit by bit) and they might eventually be able to name the concept
* SD patients remain very impaired even if they are cued right up to the end of the word
* While SA get nearly all of the names correct if they’re given enough of a cue, concept is there in their minds but they can’t retrieve them at the exact moment
Jefferies & Lampen Ralph (2007) found ventral lateral prefrontal injury, particularly in the left hemisphere can give rise to the problems controlling semantic retrieval that are related to the problems controlling retrieval that we’ve already seen
Semantic Dementia: degradation of amodal semantic representation
Semantic Aphasia: deregulated semantic retrieval; damage to control processes
How might the ventrolateral PFC play an important role in allowing us to separate semantic from episodic memory and why is this important?
helps us avoid false memories which are driven by semantic stereotypes of what we expected to see
Stampacchia et al. (2018) looked at patients who had semantic aphasia following a VLPF stroke.
Looked two tasks:
1. asked people to encode word pairs i.e. elephant goes with bear
2. test phase, sometimes we introduce semantic distractions i.e. was elephant with bear or with rhino
what did they find?
those with SA find it difficult to reject rhino because they want to choose the most semantically similar answer
- having a false memory -> rated themselves as confident with the wrong answer
- the more semantically connected the distractor was, the more confident they were in their recall
what did stampacchia et al. (2018) also find in paradigm where we paired items on a computer screen before a semantic memory judgement (asked which one was at the top), found after pairing them i.e. elephant and bucket, asked people to make semantic decisions (wanted to choose elephant and bucket when they were previously paired together)
also show false semantic association based on recent experiences
-> suggesting VLPFC helps to resolve competition that you experience between semantic and episodic memory stores -> need to know which types of representations to pay attention to, to get the appropriate behaviour out at any moment
in fMRI study, where you manipulate the difficult of making semantic decisions, what do you find?
a distributed network and large response in the VLPFC and areas of the posterior temporal cortex
what is the large scale network important for?
controlling semantic memories
what happens if the VLPFC is damaged?
people have problems controlling their semantic and episodic memory retrieval
- those who have damage to the posterior temporal cortex have similar problems
what can we suggest if those with VLPFC damage and temporal cortex damage have similar problems?
there is a distributed network of brain regions which have similar functions
what happens if you apply TMS to disrupt the VLPFC and Posterior Temporal cortex?
you disrupt the processing of hard semantic trials
- get the same level of disruption when you have to retrieve a weak semantic trial which requires more control demands compared to an easy semantic trial easy trials (only disrupt of a hard semantic trial in both brain regions)
- appears to be a large scale network, so multiple brain areas that contribute together and interact to support semantic control -> overlap between this network and regions which support control of episodic memory
what are two brain networks important for different aspects for memory retrieval
DMN and CMR
Default Mode Network
automatic aspects of retrieval from strong cues or situations where the memory has been encoded very strongly
controlled memory network
resolving competition between memories, weakly encoded episodic memories
how are the DMN and CMR linked?
CMR IS ADJACENT TO DMN (right next door) and it appears to constrain the way the DMN operates in order to help you to recover the information you need at a given moment in time
what is forgetting often produced by?
interference / trace decay
how can we control inference?
VLPFC engagement
damage to VLPFC can cause what?
problems retrieving relevant information and overcoming competition in both episodic and semantic memory
what is the new psychological perspective (cog neuroscience) about forgetting?
forgetting arises from need to control competition in retrieval (for the memories that do matter)
- selective retrieval requires suppression of competitors and that leads to forgetting
- forgetting distracters reduce need for VLPFC during retrieval
- strikes a balance between forgetting and interference - can’t avoid either completely
why is forgetting useful?
- Allows us to retrieve the information we need and reducing the metabolic demand on the brain (forgetting helps this -> don’t have to activate as much as the brain to recover information by reducing the control mechanisms need to engage)
- Brain uses 20% of energy
- Forgetting helps reducing interference
(Anderson et al. 1994) designed a retrieval induced forgetting paradigm
* People learn word pairs (i.e. drink-scotch, drink-gin)
* Then you retrieve only one bit of the information (i.e. only practice drink-scotch)
* Final retrieval is tested (have to remember the fruits and drinks you learned about)
What did they find?
- you get a difference between the memories you practice and the ones you don’t
- Practice fruit goes with orange better than drink goes with wine
- Fruit goes with banana is suppressed because it is not practiced and therefore you begin to forget information which is semantically similar to the one you practiced
- If you don’t practice either, then they don’t change
- If you selectively practice a subset of information that you’ve learned about, your memory for the bits you don’t practice goes down
ways mechanism in the brain is important for this phenomenon (by using independent cues)?
Can see the representation of banana has actually be suppressed
* When trying to not remember fruit goes with banana. and asked them what goes with monkey, the probability of them saying banana is reduced -> concept of banana is not accessible as it would have been before the selective practice
what is the neural basis of RIF effect? (Kuhl et al. 2007). People were asked to repeatedly retrieve fruit-orange. What did they find?
As they practiced that, activation VLPFC (task gets easier) reduced -> so task gets easier as you practice
* Also found suppression of concept banana, was associated with reduction in VLPFC response (greater forgetting of competing memories like banana is associated with decreased PFC for orange after practice)
^ shows how much retrieval reduced forgetting affected participants
what can we suggest from frothing the competition that they did not practice more? (Kuhl et al. 2004)
Forgetting the competitive they didn’t practice more -> more suppression of banana activatining the PFC less -> suggesting the reduction in the strength of competitiors does allow you to retrieve information more efficiently without the strong metabolic demands of memory retrieval network
can we intentionally forget? what is the view of Freuds?
Repression -> unconscious and automatic
-> memories being repressed, from your past, unconscious and influence behaviour but were pushed down by the ego in freud’s account because of their risk to our psychological wellbeing
can we intentionally forget? what is the more modern view?
suppression -> conscious process; goal-driven/intentional and related to executive control
-> more modern view is that it’s conscious - we can forget painful and damaging information but this is a conscious and goal driven process
behavioural control (cognitive neuroscience)?
Go/No-Go Task
-> the ability control actions based on goals
Go/No-Go Tasks
- measures inhibitory control over actions
- press the letter O as fast as you can, but don’t press the button if the letter is X
-> Action to press press becomes very automatic and it is quite difficult to do when the x comes
Cognitive Control
Think/No-Think Task
-> the ability to flexibly control thoughts in accordance with goals
-> stop unwanted thoughts from entering consciousness
Think/No-Think Task
measures inhibitory control over memory
-> not think about some associations and think about some
-> these kind of cog control are parallel and control related brain mechanisms
(Anderson and Green, 2001) Think/No-Think Paradigm
Participants learn word pairs in Phase 1
Phase 2- asked to selectively practice or inhibit some of these associations (i.e. steam went with train if in green! But if in red, inhibit the matching pair) -> try to think about the appearance of the letters or something - people use different strategies
* Subsequent effect of asking people suppress memories in this way
TEST PHASE: now I want you to remember the words taught at the beginning of the experiment
Or an independent probe -> giving them a semantically connected item (i.e. an insect beginning with r) - if made to suppress previously, they probably won’t say roach
What did they find?
- Compared with the baseline condition
- Think trial, items practiced were retrieved better
- Items you have practiced not thinking about are retrieved worse and less likely to be produced following the independent probe
- We do have the capacity to consciously suppress memories that we don’t want to think about -> having an impact on the accessibility of these memories later
what did Anderson et al. (2004) find for the Think-No Think Paradigm in fMRI?
PFC and more dorsal areas is on fire when we are inhibiting memories we don’t want to think about -> it’s a real cognitively demanding task
* But activation in hippocampus is reduced as we are not retrieving information -> you are able to suppress these representations
how can we understand the ideas of these brain mechanisms in PTSD and flashbacks?
- Find that flashback type memories are really type common -> vast majority of people will experience these intrusive thoughts immediately after to about 6 months after
- Some have relatively mild which resolve quite quickly
- Large group up to 35% who have really upsetting flashbacks but between 6 months and a year are resolving
- BUT we are worried about the other people, where flashbacks and disruptions don’t recover after a traumatic event -> they don’t recover and sometimes get worse over time
what can traumatic events lead to?
vivid and intrusive flashback memories
-> PTSD - trauma results in persistent anxiety often accompanied by flashbacks
Catarino et al (2015) thought there must be something about how the brain is functioning. So used Think/No-think Paradigm:
- Picture of a doll (neutral) paired with a traumatic image of people sheltering from an attack (unpleasant) -> learning the connection between the two pictures
- Then think, baseline and no-think condition
- Think: doll, have to recall traumatic image
- Baseline: no practicing
- No-think: see doll and asked not to think about the unpleasant picture in the any way you can -> very hard to do
What did they find for controls?
Controls without PTSD but still exposed to a traumatic event show the no-think effect in the way you’d expect (able to successfully not think about them following instructions and show poorer recollection later) [difficulty with inhibiting memories in a deliberate manner]
Catarino et al (2015) thought there must be something about how the brain is functioning. So used Think/No-think Paradigm: What did they find for those with PTSD?
Those with PTSD don’t show the suppression effect -> vulnerability to PTSD, less good at suppressing memories they don’t want to think about
what is the VLPFC important for and why?
memory control
- helps to select relevant information and overcome competition to both episodic and semantic retrieval
damage to PFC is linked to what?
poor source memory and confabulation
VLPFC damage causes..
deficits of controlled semantic retrieval in semantic aphasia (dissociates from semantic dementia)
can forgetting be positive?
yes. reduces interference and demands on the VLPFC
can forgetting be intentional?
inhibition of unwanted memories is supported by VLPFC
what is intentional forgetting impaired in?
PTSD
