Lecture 3 - Episodic Memory and Forgetting

episodic memory

part of LTM, memories for specific episodes or events that are tied to the time and place at which the information was learned, requires knowledge of the self and consciousness

hippocampus

binding content to context

what is stored in the encoding phase?

the position of information, levels of representation and context

how can we optimize encoding during the retention interval and testing phase?

through learning orders, contextual variability, testing effect, organization, distinctiveness and elaboration

primacy effect

people better remember stuff that comes at the beginning of the list in LTM because the information is novel, since its the first time you see that info your brain is excited and wants to encode it

recency effect

in LTM this is much more pronounced compared to the primacy effect, idea that people remember the stuff that comes at the end of the list better due to decay, newest info so less likely to be forgotten

levels of representation

information is processed and stored at multiple levels

surface form (levels of representation)

verbatim word for word, whatever the sentence in that piece of news is that’s what you remember

textbase (levels of representation)

abstract representation of the meaning, remember key points and fill in the rest, getting the gist of the info

mental model

mental simulation of the described events, most abstract type of representation

levels of representation experiment

participants read paragraph and in it there is a piece of memory they are trying to target, they are presented with different sentences similar to the target one and the researchers looked at old and new recognition with signal detection theory (either exact sentence, paraphrase or interference where idea is kinda there)

context appropriate condition of the levels of representation experiment

related to the target sentence

unrelated condition of the levels of representation experiment

not related to the target sentence

what were the results of the levels of representation experiment?

participants were good at calling the surface level retrieval and could differentiate old from new paraphrased target sentence, surface level information is good at the beginning but disappears after awhile

what is the textbase information in the levels of representation experiment?

based on the paraphrased and interference information, a bit more abstract then the surface level info

what was the model in the levels of representation experiment?

based on inferences, if it is context appropriate, in the study it was remembered for the longest amount of time cause higher level of representation and therefore encoding

what is the key takeaway of the levels of representation experiment?

mental model represented information is maintained the longest in memory

context experiment

participants studied a list of words in different contexts and had to do recall, researchers varied if the context at encoding matched the context at retrieval

context experiment results

memory was best when retrieval and encoding contexts were the same (ex: if participants studied on land and were test on land they did better than if studied underwater and tested on land or vice versa)

encoding specificity

memories are linked to the context in which they are created

learning order experiment

had a list of word pairs and each of them appeared twice and the researchers wanted to see if there was a difference between distributed or mass learning, list was a bunch of pairs back to back that would reappear close or further along the list

learning order experiment results

participants were able to recall more words in the distributed practice rather than mass practice because in the mass condition there wasn’t enough time for consolidation to happen cause not enough rehearsal time + more variability in distributed compared to mass learning

study phase retrieval theory

repetition serves as a retrieval cue to reactivate and then strengthen the representation of the prior experience which might be why distributed learning is better than mass

what is the mechanism for distributed vs mass learning

when learning is distributed doing it in the same context enhances study phase retrieval but if learning is mass varying the context you learn is enhances study phase retrieval

what is the caveat for context in improving memory retention?

varying the context will help you remember things better in long term memory but it is dependent on timing

varying context experiment

face name pairs were superimposed on videos in cued recall design and they received feedback then later there was a final test where it was a cued recall with a new or old video, researchers manipulated consistency vs variability in the learning phase

varying context experiment results

in the constant condition they were way better when recall was immediate compared to the varied condition but when tested 2 days later the varied condition did much better than constant so variable context was better in the long term

what might explain the results of the varying context experiment?

studying across multiple varied context allowed more routes to retrieval because less context specific the info is the more accessible it is across a range of situations

testing effect experiment

people read a paragraph then read it again and either do free recall or cued recall and at the final recall phase the researchers were looking at the retention interval

testing effect experiment results

both conditions did about the same in the beginning but over time both dropped however the study test (cued recall) condition helped remember things better than the study-study(free recall) condition

what is the reasoning behind the testing effect experiment results?

could be the recall process, trying to come up with info from that paragraph might be leading to elaborative rehearsal of that material, might be reducing the rate of forgetting, adds contextual variability (fill in the blanks), reduces proactive interference 

organization experiment

recall was better if the word list presented is organized in hierarchies for example rather than random lists because it mimics chunking in long term memory

distinctiveness experiment

memory is better for distinct elements in a set of information, for instance if in a word list there was one red list while the rest were black you would remember the red one better cause it stands out

von restorff effect/isolation effect

distinct item can be isolated from contrasting context for example studying tulip among vehicles

bizarre imagery effect

if you imagine things all together creating a very weird image will help with memory

relational processing

sort into categories

item specific processing

rate item for pleasantness

material appropriate processing

memory is better if the type of learning emphasizes the information for which memory is likely to be weak

many per category condition of the organization vs distinctiveness

processing is relational, category sorting where distinct facts are already highlighted

few per category condition of organization vs distinctiveness

processing is item specific, highlight distinct features

what are the results of the organization vs distinctiveness experiment?

in the many per category condition, you can recall more words if the encoding is item-specific but if the info that is shared is already highlighted what you do instead is whatever memory is weak for that’s what you’re gonna focus on but in the few per category condition sorting is helping you remember better

feature cue

provide with the pairs and provide only one item and ask them to recall something

retrieval cues experiment

participants were presented with 48 words from 12 different categories where the 1st group was tasked with free recall while the other used cued recall with the category names acting as the cues

retrieval cues experiment results

the cued recall group recalled 30 out of the 48 words while the free recall group recalled 19 out of the 48 words

associative strength

the best retrieval cue is the cue that occurred most frequently with the item in the past

encoding specificity

the best cue is that which reinstates the original context, if the cue was present at encoding and retrieval memory performance is better

encoding specificity vs associative strength experiment

in one condition words are presented back to back and in the other a weak cue was presented (sky-dark), in the retrieval phase it was either free recall weak cued recall or strong cue (light-dark), associative strenght is related to the strong cue and weak cue is associated with encoding specificity

encoding specificity vs associative strength experiment results

 they did better if they were presented with no cue if there was no cue at encoding, if they were presented at weak cue at encoding and participants did way better if the same weak cue was presented at the testing phase (match context at encoding and retrieval), shows that encoding specificity is more important than associative strength

transfer appropriate processing

memory is better when retrieval conditions resemble encoding conditions

transfer appropriate processing experiment

need to say if words make sense in the contexts, in the test phase was rather a recognition test or provided with rhyming recognition text (mimicked the same thing again) and researchers looked at proportion correct for the standard vs rhyming one

standard condition of the transfer appropriate processing experiment

if they filled the sentences, better memory for the rhyming one than the meaning one

rhyming condition of the transfer appropriate processing experiment

did better if they did a rhyming test at encoding and a rhyming test at retrieval 

what was the key takeaway of the transfer appropriate processing experiment?

Deep encoding leads to better memory but shallow encoding leads to better memory if tested in surface based tasks

state dependent learning

idea that memory is better when person’s physiological state at retrieval matches their state at encoding

state dependent learning experiment

there were four conditions, the first was the participant was sober at encoding and retrieval, the second was the participant was sober at encoding and intoxicated at retrieval and the third was the participant was intoxicated at encoding and sober at retrieval, lastly, the participant was intoxicated at encoding and retrieval

state dependent learning experiment results

unexpectedly the condition where the participants were intoxicated at encoding and retrieval made less errors than when the participant was intoxicated at encoding and sober at retrieval showing that memory is better when person’s physiological state at retrieval matches their state at encoding

mood dependent learning

memory is better when retrieval mood matches mood at encoding

part set cuing effect

cue causing poorer memory retrieval relates to the vulnerability of retrieval plans cause if for example you are trying to remember line 2 subway stations and someone says, st George first you’ll have more difficulty than cued recall cause you would want to retrieve it in the order you learned it in

retrospective memory

memory for the past

prospective memory

remembering to do things in the future

event based prospective memory

remembering to give some message to someone when you see them

time based prospective memory

remembering to call mom on mother’s day

episodic future thinking

imaging or planning for the future, form of mental time travel, has similar neural substrates as thinking about the past

constructive episodic simulation hypothesis

suggests that we use memories of previous episodes to construct a representation of the future, easier to construct thoughts of the future when they conform to our past experiences

what are the seven sins of memory

absentmindedness, transience, blocking, misattribution, suggestibility, bias and persistence

absentmindedness

forgetting due to lapses of attention during encoding that results in subsequent failure to remember information, common source of everyday memory failures

what processing styles might result in absentmindedness?

shallow processing and automatic actions without attention

absentmindedness experiment

the experimenter change in the middle of interviewing someone when an object obstructing the participants vision passed by and the participant didnt notice the person interviewing them changed (change blindness), an example of absent-mindedness at encoding

how does absentmindedness take place at encoding?

not being attentive to the situation you’re in, divided attention for instance taking a photograph of an event rather than being present in the moment would worsen your memory of it, impacts availability of the memory

how does absentmindedness take place at retrieval?

prospective memory failure, impacts the accessibility of the memory

prospective memory failure

forgetting to carry out a function you planned like setting an alarm but later forgot what it was for or got distracted

transience

memory for facts and events are lost from storage over time, similar to the idea of decay, no longer available due to lack of use

law of disuse

memories themselves decay over time if not used, controversial take as some believe the passage of time causes nothing on its own but it must be correlated with something that does cause forgetting

what is the new theory of disuse?

the memories themselves remain but the more time passes without retrieval the more difficult it becomes to access

storage

quality of encoded representation

retrieval

ease with which information can be retrieved

blocking

trouble accessing a memory because other memories get in the way, information is temporarily inaccessible at retrieval, might be available but cant be accessed, happens for episodic and semantic information

what is the evidence for blocking?

influence of retrieval cues (performance is better for cued than free recall), interference (competition between memory traces, the primary cause of forgetting in LTM)

proactive interference

occurs when old knowledge results in increased forgetting of new knowledge

• ex: if you studied french then Spanish and take a Spanish test the old knowledge of french gets in the way leading to worse performance on the test

how does proactive interference work over a long period of time in LTM?

it builds up over time and can be released with information that is new/distinct from old knowledge and greater the amount of related information produces more and more proactive interference

retroactive interference

occurs when new knowledge makes it difficult to remember old knowledge

proactive interference experiment

the participants had to study and recall 4 different lists, the first 3 were all different types of fruits so when participants were asked to recall them the old information from the previous lists for instance the first interfered with the 3rd list cause they were similar, but the 4th list differed across groups and the build up of proactive interference was released, professions were remembered the most cause if was the most different from the fruits

why do we see retroactive interference?

unlearning of prior associations, retroactive interference can be reduced or eliminated later and the meaning of memories are still available but not accessible so it could be that learning new information disrupts the retrieval plan

associative interference

number of associations with a concept can cause interference because related concepts can compete during retrieval

fan effect

the more concepts you acquire the more difficult it is to retrieve and fan refers to the number of associations correlated with the concept

why do we observe the fan effect?

the more associations there are with an item, the greater the interference for that item (the other associations we have get in the way), explains why cues with high overload are less useful

fan effect experiment

fan level of locations is calculated and the fan level is determined by amount of associations made (depends on person or location), researchers looked at level of fan and response time to recognition to the sentences

fan effect experiment results

• If fan level was at 1 (one association): the response time was very quick

• If fan level is increasing (more associations with the concept): response time increases, takes longer to figure them out and respond

what is the key takeaway from the fan effect experiment?

the more associations there are with a particular memory trace the longer it takes to verify the memory

how do we control or reduce interference?

by inhibition, actively reducing the activation of interfering information but inhibiting related memories can also cause forgetting

what are the negative consequences of inhibition?

part set cuing and retrieval induced forgetting

part set cuing

alternative explanation to retrieval plan disruption, providing some cues to inhibit other or unrecalled memories

retrieval induced forgetting

remembering one thing makes it harder to remember other info

persistence

remembering a fact or event that one would prefer to forget, failure to forget

directed forgetting

explicitly tell participants to forget some items and remember some others, in the experiment people were able to forget things they were told to forget showing we have some control over what you want to forget

persistence experiment

control participants had no problem forgetting the words but with PTSD patients they were not able to forget trauma related words cause they couldnt control what they could remember and what they could forget

retraction

previously learned information is later marked to be incorrect and should be forgotten

continued influence effect

continue to believe misinformation even after it is corrected

event boundary

meaningful change in the ongoing event such as change in location, jump in time or change in activity, leads to mental representation of the new event so old memory is moved out of working memory

how could event boundary disrupt memory

when people walk from one room to another, their memory of the objects is worse than if they stayed in the same room