1. Long-Term Memory

2. The Traditional View of Long-Term Memory
Capacity
Analogy:
“Scrapbook” or “treasure chest”
What type of info is stored?
Some think LTM is unlimited
Modal model of memory (modal refers to idea that there are modules, or component – something that has a particular function)
Capacity – amount of information a module can store
LTM often thought of as a scrapbook or treasure chest, something that stores info,
What type of info is stored?
What do you keep in your long term memory? Word meanings, facts like math, history, other info you have learned, names, faces, directions, facts about people (birthdays, physical characteristics, favorite goods), how to get a MARTA pass, how to read a train schedule, how to ask for a date, …
So some think LTM is unlimited
usually thought that memories will not change
How is this different from STM?
capacity shorter
Also, one natural question is, where does memory go when you cannot find it?
Has it disappeared? Lost connection point?

3. STM LTM However…. I just sat down. I had lunch.
I got my drivers licence.
I walked to class.
Long-term memory has differences, some more detailed than others
Years Hours Minutes Now

4. The Traditional View of Long-Term Memory
Coding
One explanation for errors of retrieval
Problems with encoding
Coding – how we represent info in brain
One explanation for errors of retrieval - Problems with encoding
Ex. of problems with encoding
semantic confusions: experiment
list of words, sounded similar or did not (mad, map, man: pen, day, rig)
or ones that similar meaning or no similar meaning (sim - huge, big; diff - foul, deep)
20-min interval, worked on another task (so prevent rehearsing, and make sure from long-term mem, not STM)
R: similar sounds, no effect; similar meanings, harder to learn
generalization: STM affected by similar sounds, LTM by similar meanings
We’ll talk later about

5. Long-Term Potentiation

6. The Traditional View of Long-Term Memory
Retention Duration
Recognition:
Recall:
Experiment 1: recall of Spanish
Experiment 2: recall of spatial layout of a city
Retention Duration: How long can info be stored in LTM? Years?
Difference between…
Recognition: remembering if info presented previously
Recall: remembering details of info with few cues
Experiment: students in Spanish class, some currently there, some hadn’t taken class for 1-50 years;
sharp decline for first 3-6 years, leveling off of forgetting, then final decline at years
So, a lot of info retained even though not rehearsed or used
Experiment: recall of spatial layout of a city:
1. recall of street names, building names,
2. location of streets/buildings on map (no names given)
3. matching list of names to map (names given)
recognition: retrieval of info, by making a choice between different possibilities (matching)
recall: retrieval of info with few aids (free recall)
Result: streets learned at slow, steady rate; but also forgotten very quickly, and most of info lost after 10 years
buildings learned quickly, but remembered for a long time (40% remembered after 46 years)
most important variable seemed to be how important items were to Ss (buildings on campus learned faster than streets where students not as likely to go)
Conclusion: info can be retained for long periods, but most important factor may be how important info was to person, and how well they learned it to begin with

7. The Traditional View of Long-Term Memory
Forgetting
Ebbinghaus’s forgetting curve
Forgetting – being unable to recall part or all of desired info
There are degrees of “knowing” – recognition vs. recall, how many details, knowing part of story but forgetting punchline
Ebbinghaus’s forgetting curve
Created lists of nonsense syllables (no meaning in English, like rur, meb), memorized them, later tried to recall them
Varied lots of different things, like how long he spend memorizing, time between memorizing and recall, length of words, etc.
Most well-known finding – forgetting curve
Graph – percent remembered, amount of time between memorizing and recall
Conclusion – we forget stuff very rapidly at first, then the amount of forgetting levels off and we retain what was remembered for a long time
NOT linear

8. The Traditional View of Long-Term Memory
Forgetting
A learning task: Paired-associates learning
Why do we loose info, if LTM supposedly keeps so much?
Same question as with STM: Decay or Interference ?
Tests on interference often use paired-associates learning task
A learning task: Paired-associates learning
Ss hear list of word pairs, then shown first word, asked to recall pair associated with it

9. The Traditional View of Long-Term Memory
Forgetting
Interference
Proactive interference
Retroactive interference
Two types of Interference
Proactive interference: previous learning makes later learning more difficult
for paired test, exp group, learn list A-B (dog-stairs, tree-cup), learn list A-C (dog-clock, tree-car), test for list A-C
control group first unrelated activity, then learn A-C, then test A-C
recalling info from second list is harder, b/c of first list learned
real-life example – learn French-English vocal first, then German-English,
German will be harder to learn, compared to control group (control group will learn better)
and exps have shown that the more experience you have with first list, the harder it will be to recall second list
Retroactive interference: later learning makes earlier learning more difficult to recall
for paired test, exp group, learn list A-B (dog-stairs, tree-cup), learn list A-C (dog-clock, tree-car), test for list A-B
control group first learn A-B, then unrelated activity, then test A-B
control group will learn better

10. The Traditional View of Long-Term Memory
How does interference work?
Retrieval Cue
Fan effect
How does interference work?
Retrieval Cue – one item becomes associated with another, first item can be used to help remember second item
Problem occurs when same retrieval cue gets associated with more than one other item
Competition b/t two items for same cue
Parking car example – if park in lot once, easy to remember where you parked,
if parked several times, harder to remember
Fan effect – as you learn more about a topic, creating more associations with same cue, it will be harder (take longer) to pull out a specific fact from your knowledge base (harder to pull out the ace of spades from a hand of 15 cards than a hand of 2 cards)
Sometimes good to forget unneeded info so that it will not interfere with upcoming learning
Demo: read list
Apple, desk, shoe, sofa, plum
Chair, cherry, coat, lamp, pants
Grape, hat, melon, table, gloves
Write down as many as you can remember
What did you remember? What order did you remember them in? are any of your answers grouped into categories?
Fruit, furniture, clothing

11. The Traditional View of Long-Term Memory
Retrieval of Information
Mnemonics
Categorization
Something similar in another experiment where people also not told that words could be in categories
Experiment by Tulving and Pearlstone: gave people a list of words to remember, all words were from specific categories (birds: pigeon, sparrow; furniture: chair, dresser; professions: engineer, lawyer), but participants were not told this when they first got the list
Two groups – one asked to recall list in free recall condition – no cues
Other group given categories of words
No cues – 40% of words recalled; categories – 75% of words recalled
Categories were retrieval cues
Categorization: easier to remember things if you can group them into categories
for studying, make lists of logical categories
This is an example of a mnemonic: techniques to improve memory

12. The Traditional View of Long-Term Memory
Retrieval of Information
Mnemonics
Encoding specificity
State-dependent memory
Environment (context effect)
1. Learn List
Encoding specificity: context of learning important, in other words, when we learn information, we pay attention to our surroundings, and associated the learned info with that environment – the environment itself becomes a retrieval cue
think of hooks being attached to info to be remembered at the time of encoding
State-dependent memory – refers to the condition our body is in during learning (location, feelings, drunk, hyped up, tired, )
Environment (context effect): scuba experiment, two groups, one - scuba divers learned list of words underunder, two – list learned on land
Test: each group divided into half; half recall list on land, half underwater
Results: tended to remember words best if, when asked to recall them, they were where they learned them
Maybe parts of the environment (location) itself become retrieval cues
Have any of you ever noticed that when you are taking a test it might seem abnormally quiet, that it is hard to get yourself to remember info?
What might be happening is that you normally learn the info in a noisy environment, music playing, friends talking, maybe in a place where other people are talking in the background – then you get to the test environment where things are different – there is almost no noise
2. Testing
Location
Underwater
Land
Underwater
Land

13. The Traditional View of Long-Term Memory
Retrieval of Information
Mnemonics
Encoding specificity
State-dependent memory
Physiological state
Mood
Same sort of thing might happen with physiological state or mood
Physiological state – say you are cramming for a test and drink a lot of caffeine to stay awake (or other stimulants);
If you take the test without these stimulants in your system, you will have a harder time remembering the info than if you still had stimulants in your system
Same thing for being drunk (not to say you should get drunk when learning and get drunk when recalling the info!)
Mood – same for mood – if you learn info when you are happy, you will be able to remember it better when you are happy

14. The Traditional View of Long-Term Memory
Retrieval of Information
Mnemonics
Encoding specificity
Spacing effect
Encoding variability
Spacing effect: cramming is not as efficient! If you repeat learning twice in short period of time, it is no more effective than learning it just once, but if you space two learning sessions apart, it is twice as effective (in other words, you will tend to remember twice as much)
Why? one theory – encoding variability – learning at different times allows you to learn in different contexts, providing more hooks
more hooks can be attached to same info with repeated sessions, so it becomes more likely that info will be recalled
Cue overload: best when cue is distinctive and not related to other targets (remember C for cannon, categorization; not as effective if another concept begins with C, like cue) (related to fan effect – increased retrieval time if too many things associated with a cue)
Amount of processing: black without another word, white will make black come back more easily to memory
black with word train, more encoding takes place, so seeing train makes black come even faster than the word white

15. The Traditional View of Long-Term Memory
Retrieval of Information
Mnemonics
Encoding specificity
Cue overload
Amount of processing (black:white; black:train)
Cue overload: best when cue is distinctive and not related to other targets (remember C for categorization, with an experiment by someone named Campbell; not as effective if another concept begins with C, like cue, or there is another scientists with a C name like Collins)
You are going to start getting confused
(related to fan effect – increased retrieval time if too many things associated with a cue)
Amount of processing: black without another word, white will make black come back more easily to memory
black with word train, more encoding takes place, so seeing train makes black come even faster than the word white

16. The Levels-of Processing View
Basic assumption: Depth of Analysis
Superficial or shallow encoding
Deeper or semantic encoding
Instead of length that info stored, and types of different storage,
Another view is that it is all one memory, the key being how much info was processed when first being encoding
(i.e., there are different kinds of cognitive processing)
Superficial or shallow encoding – rehearsal, repetition
Deeper or semantic encoding

17. The Levels-of Processing View
Basic assumption: Depth of Analysis
Experiment:
Capital letters?
Rhyming?
Fit into sentence?
BABY monkey
Experiment: Ss given a question, then a word
three types of questions
Is the word in capital letters? (physical processing)
Does it rhyme with ______? – does the word rhyme with hat? (acoustic processing)
Fit into sentence? – does the word fit into the sentence – The girl placed the _____ on the table. (semantic processing)
R: semantic had better recall
Also, semantic processing can have different levels: simple sentence vs. complex sentence, easier to recall if had complex sentence
elaboration
Importance of this new theory – got people thinking about HOW info encoded
cat fish
plate ocean

18. Short-Term Memory vs. Long-Term Memory
Time info stored
Capacity
Coding (form of storage)
Ways info retained or forgotten
Ways info retrieved
Time info stored
ST – a few mintues LT – forever?
Capacity
ST – LT - infinite
Coding (form of storage)
ST – auditory LT - semantic
Ways info retained or forgotten
ST – rehearsed LT – never really forgotten?
Both may be lost by inference
Ways info retrieved
ST – serial, exhaustive search LT – retrieval cues, Mnemonics

19. Suggestions for Studying
Elaborate
Organize
Associate
Take breaks
Match learning and testing conditions
Suggestions for studying:
Elaborate -
Organize
Associate
Take breaks
Match learning and testing conditions

20. Question:
Do lab experiments really tell us anything about how memory works in real life?
Some people say no,
Because memory can be distorted, which lab doesn’t tell us much about
Ex. telling a story/ remembering a story
my story of barn, dogs, parents leaving
theory that people distort info (unintentionally) to make it make more sense

21. Reconstructive Nature of Memory
Autobiographical Memory
Barlett gave Ss a story, asked later (sometimes years later) to repeat story
He was able to show that details of the story changed, “foggy and calm” night changed to “dark and stormy” night
Here memory is not unchanging, it is an active process
Auto: memory for events that happen to person doing remembering
Trying to figure out when something happened in past, we often use logic
Errors: Sometimes, when we try to remember something, we use logic, and deduce what must have happened, not really remembering, just reconstructing
two kinds of unremembered events
unrecalled items – just not remembered
unable to distinguish one event from another similar event
When try to make people give you a specific event, people often give a summarized (occurred more than once) or extended event
But, overall, people actually pretty good at remembering things, with more distinct events easier to remember

22. Reconstructive Nature of Memory
Flashbulb Memories
People often remember own circumstances when recalling an important event not directly related to them
Why? possible theories
link to history  tell own story
strong emotions  (stronger the emotion, stronger the memory)
But these memories can be distorted over time too

23. Reconstructive Nature of Memory
Eyewitness Memory
Elizabeth Loftus experiments
Exp 1: slides of a car hitting a ped; one set of slides car came to stop sign, other set a yield sign
later Ss asked questions that were either in sync with what they saw, or out of sync
Ex. Did another car pass the red Datsun while it was stopped at the yield sign?
Other unrelated questions asked, did unrelated activity,
20 mins later asked if they recognized photo of car
R: when question not in sync, people recognized picture only 41% the time (as compared to 75% when in sync)
Memories are changeable depending on what happens afterward
Exp 2: viewed film, then asked misleading questions,
how fast was the car going when it passed the barn on the country road? 17% remembered barn
how fast was the car going on the country road? 3%

24. Reconstructive Nature of Memory
The Recovered/False Memory Debate
Exp. interviewed relatives/friends of a person who knew them well when they were 4-6 years old, then created a false event
showed S four stories, three real, one false, 25% reported false memory true
Explanation: idea of false memory goes into your head, you forget that it was false, we use things that really happened, put them into false memory,
However, some scientists point out that just because false memories can be made this way does not mean that they are in reality
There are limits to what a person can be made to falsely remember – the memory has to be something that they have some background reference for,
Autobiographical memory is not a video camera, does not preserve info perfectly for later retrieval

25. Neurological Studies H.M. case
HM- had parts of hippocampus, amygdala, and adjacent areas (rhinal cortex)
Result; He could remember semantic info, and info from several years before surgery,
But he could not form new memories
And could not remember things just a few years before surgery
H.M. case

26. Amnesia Anterograde Amnesia Affects LTM but not working memory
Affects memory regardless of modality
Spares general knowledge
Spares skilled performance
Although skilled performance depends on context (hyperspecific memory)

27. Amnesia Retrograde Amnesia Temporal extent varies among patients
Most recent memories are most likely to be lost
Spares “overlearned” information
Spares skilled learning