Overheads from Memory, PSY 383

What is Memory?

Metamemory: Knowledge and awareness of one's own memory processes and abilities

Egypt 4000 B.C. Thoth: god of learning, memory and wisdom

Greece 1000 B.C. Mnemosyn: goddess of memory

First text on improving memory: Dialexis 400 B.C.

The three most important aspects of memory are: remembering to pay attention, to rehearse and to use a formal mnemonic device (the method of loci)


First memory theorists:

Plato (427 - 347 B. C.) All essential truths are stored in memory and learning is the process of recollecting these truths.


* Wax tablet model: Memory is like impressions formed on a wax tablet


* Aviary model: Memories are like different species of birds.


* Scribe model: A personal scribe who records the events of ourlives

Aristotle ( 384-322 B. C.)

Recollection is based on associations

Associations are connections of mental events

Three laws of association: events tend to be associated when


* they occur together in time or space (law of contiguity)


* they are similar (law of similarity)


* they contrast with each other (law of contrast)

Cicero (106-43 B.C.) and Quitiliian (40-96 A.D.)

Emphasis on the role of memory in oratory:

The order of remembering is important

Physical and mental exercise helps

Augustine (354-430 A. D.)

Remembering is like exploring caverns.

Illusions of memory are possible when "we fancy we remember as though we had done or seen it, when we never did or saw at all".


Hermann Ebbinghaus (1850-1909)

Scientific approach to memory

studies his own learning of nonsense syllables

buv, weq, rol

Controlled: time of study, interval between study and recall

Measured amount of retraining needed based on:

interval, size of list, amount of studying

General theoretical apporaches to memory:

Biological: factors involved in the physiological structure and formation, storage and retrieval of memories

Associationism/connectionism:

Information-processing/cognitive: Memory viewed in terms of how information is processed.

Multimodal: Many different influences on memory (strategies, type of material, motivation, emotions, physiological factors, age)


Methods in Memory

Naturalistic Observation

Observing information in a non-experimental setting

When do people forget things? What factors are involved?


* Data from naturalistic observation.


* Rich, but hard to analyze.

Controlled experiments

Goal: test an empirical hypothesis

Hypothesis: Students will remember more if I say everything twice Students will remember more if I say everything twice

Independent Variable: Variable that is manipulated to test the hypothesis.

Dependent Variable: Variable representing the behavior we want to measure

Control Variables: Other variables we need to control in order to see the effect of the independent variable

Subjects: Who is going to participate in the experiment?

Analysis: How do we know if there are differences between the manipulation and the control group?


Measuring memory performance

How do we go about studying memory?


* Registration
: transforming information into a form that can be retained


* Retention
: storage of information


* Remembering
: Retrieval of the information

Implicit vs. explicit remembering


* Explicit: information explicitly remembered


* Implicit remembering: No recall, but still influences performance


Independent variables in Memory research

Organismic variables

The relativiely permanent characteristics of a person that affect memory performance

(ability to pay attention, motivation, intelligence, etc.)

Antecedent variables

Variables that alter a person's typical organismic level

(drugs, external motivators, sleep)

Task variables

1) Instructional variables

Implicit or explicit instructions given to the subject

(study hard, use a mnemonic)

2) Presentational variables

How the stimuli are presented

(Time of presentation, presented verbally or written)

3) Stimulus variables

Differences in stimuli used

Familiarity (green vs. chartreuse)

Frequency of occurence (home vs. abode)

Concreteness (dog vs. faith)

Other features: shock, meaningfullness

4) Context

place, mood can play a role in encoding and recall


Measuring memory performance

How much/well do you remember?

Primary measures

Amount of information recalled

How likely is it that a subject guessed?

Secondary measures

Quality of information recalled

Speed of recall: reaction time (RT)

Evaluating memory

Serial learning: Learning to recall information in a particular order

Free recall learning: learning to recall information in any order

Cued recall: Recall given a cue

Paired-Associate Learning: Learn stimulus + response

dog-cat dog-plant dog-zik duck-pato duck-canard

Tests

Recognition: Yes/no Did you hear bed?

Which did hear?

street

pencil

toe

book


Study test procedure

Study half the list:

Then see the whole list and decide if you way it or not


On the list

Not on the list
Subject claimed seeing it
Hit
False alarm
Subject claimed not seeing it
Miss
Correct rejection

Testing implicit memory

Amnesic subjects don't recall words from the list, but if you do a word completion

Fill in the rest of the word Ch____



The Sensory store

Processes incoming information from the environment


* Individual sensory stores for each sense


* Information retained for a short duration

The visual sensory store (Iconic memory)

Experiments by Sperling (1960)

X M R K

C N J P

V F L B

The partial report technique

Auditory sensory store (echoic memory)

Experiment by Darwin, Turvey & Crowder (1972)

3 digits or letters auditorally presented to each ear and center at the same time


What is the use of the sensory store?

1) It maintains information long enough so that we can do additional processing to it.

2) Gives us time to choose what sensory information to give more attention to

3) Keeps information active long enough to get us to notice it

4) Provides perception of continuity in our environment


Working memory or short term memory (STM)

STM used to describe the fact that it holds information for a short time, while working memory refers to the processing capacity.

STM works as a temporary holding place for intermediate decisions.

Limited in size.

Chunking

Limited Time

Code: Acoustic

Working memory: there is a limited amount of processing capacity that you can use as you perform a problem

Brown-Peterson Distractor task

Task:

1) give 3 random letters and a 3 digit number.

2) Have subjects count backwards from the number by threes for X seconds

3) Recall

L W C 455


Brown-Peterson Distractor task Name_______________

Trial # Recall Number correct Delay

1




2



3



4



5



6



7



8



9



10



11



12



13



14



15



16



Delay 3 Total # correct: ______

Delay 6 Total # correct: ______

Delay 9 Total # correct: ______

Delay 12 Total # correct: ______

Delay 3 Total % correct(number correct/12): ______

Delay 6 Total % correct (number correct/12): ______

Delay 9 Total % correct (number correct/12): ______

Delay 12 Total %correct(number correct/12): ______

Brown-Peterson Distractor task

Trial # Recall Number correct Delay

1

T C X 584

3
2
V J B 832

6
3
L Q Z 387

9
4
P R H 256

12
5
M F S 492

3
6
Y Z P 217

6
7
W B T 456

9
8
G D V 782

12
9
K H W 353

3
10
Z Y W 587

6
11
T P G 548

9
12
C F S 772

12
13
R X W 232

3
14
N T B 663

6
15
H G D 924

9
16
F C L 345

12
Delay 3 Total # correct: ______

Delay 6 Total # correct: ______

Delay 9 Total # correct: ______

Delay 12 Total # correct: ______

Delay 3 Total % correct(number correct/12): ______

Delay 6 Total % correct (number correct/12): ______

Delay 9 Total % correct (number correct/12): ______

Delay 12 Total %correct(number correct/12): ______


Retrieval from STM

Sternberg Serial Scanning task

Example:

1) Memory set presented: 4, 8, 2, 5, 9

2) Subject encodes set into STM

3) Probe item presented: 5

4) Subject presses yes or no key

Implications:

1) We sequentially compare probe items

Evidence from the time increasing as a function of the number of items

2) We use exhaustive serial search.

Evidence in that there is no difference between the Yes and No responses, although, subjects could terminate the search once they see a Yes response.


Long term memory

Knowledge of: how to do things, things we have learned, grammar rules, personal memories.

All knowledge that is not active.

Information that becomes active is retrieved from LTM and put in STM.

Almost everything we learn is first processed in STM and some of it is put into LTM

Demonstration: The Serial Position Curve


Explanation of the serial position curve:

Recency effect, Primacy Effect

Recency effect: Last items heard/seen are recalled well

Primacy effect: First items heard/senn are recalled well

Why?


* Recency effect: Due to STM


* Primacy effect: Due to LTM

Does this mean there is really a STM/LTM distinction?

There may be a single mechanism that could explain both STM and LTM

Some evidence

Evidence from Amnesia:

Evidence in Anterograde amnesia for patients with LTM intact, but inability to consolidate new information from STM into LTM.

Some patients have defective STM, but fine LTM.


The structure and representation of LTM

Semantic Networks

Words can be represented as an interconnected network of sense relations


* Each word is a particular node


* Connections among nodes represent semantic relationships

How do these pieces of semantic information relate to each other?

Semantic verification task

An A is a B

An apple is a fruit

A robin is a bird

A robin is an animal

A dog has teeth

A fish has gills

A fish has feathers

An apple has teeth

NMSU is in New Mexico

Harvard is in California

Use time on verification tasks to map out the structure of the lexicon.


Models of the Semantic Memory

Feature comparison model (Smith, Shoben and Rips (1974)

Information (concepts) are stored as a set of features:


* Defining features (those features that it must have)


* Characteristic feature (those feature that it often has)

What are the features of a bird?

Comparing concepts through comparing features

Is a robin more like a swallow or or a penguin?


Collins and Quillian Hierarchical Network model

Lexical entries stored in a hierarchy, with features attached to the lexical entries

Representation permits cognitive economy

Testing the model

Sentence Verification time

Robins eat worms 1310 msecs

Robins have feathers 1380 msecs

Robins have skin 1470 msecs

A category size effect: Subjects do an intersection search

Problems with Collins and Quillian model

1) Effect may be due to frequency of association

2) Assumption that all lexical entries at the same level are equal

The Typicality Effect

Which is a more typical bird? Ostrich or Robin.

A whale is a fish vs. A horse is a fish

Major conclusions of the model:

1) If a fact about a concept is frequently encountered, it will be stored with that concept even if it could be inferred from a more distant concept.

2) The more frequently encountered a fact about a concept is, the more strongly that fact will be associated with the concept. And the more strongly associated with a concept facts are, the more rapidly they are verified.

3) Verifying facts that are not directly stored with a concept but that must be inferred takes a relatively long time.


Spreading Activation Models (Collins & Loftus)


* Words represented in lexicon as a network of relationships


* Organization is a web of interconnected nodes in which connections can represent:

categorical relations

degree of association

typicality

Retrieval of information


* Spreading activation


* Limited amount of activation to spread


* Verification times depend on closeness of two concepts in a network

Context effect in spreading activation models

Present either: Murder is a crime or Libel is a crime

Then get verification time for Robbery is a crime

Subjects faster when they see Murder than Libel. Why?

Advantages of Collins and Loftus model


* Recognizes diversity of information in a semantic network


* Captures complexity of our semantic representation


* Consistent with results from priming studies


Lexical Access

What factors are involved in retrieving information from LTM (the lexicon)?

Semantic Priming

Meyer & Schvaneveldt (1971) Lexical Decision Task

Prime Target Time

Nurse Butter 940 msecs

Bread Butter 855 msecs

Evidence for associative spreading activation

Ratcliff and McKoon (1981)

Subjects study and memorize The doctor hated the book

Task: "Was this word from the sentence you memorized?"

Prime Target Time

None Book 667 msecs

Doctor Book 624 msecs


Memory review:


Sensory store

STM
LTM
Duration
500-2000ms
<30 Sec
minutes to years
Format (code)
Literal copy
Acoustic
Semantic
Capacity
medium (9 visual items)
small (~5-7 chunks)
unlimited
Rehearsal process
not possible
maintenance
elaborative
Retrieval process

Serial
Parallel


Encoding: How do we store information into LTM?

Explicit vs. incidental learning

Selective processing of information

We do not encode all information from stimuli

Effects of expectations

We encode information based on how we believe we will need to use it?

Effects of expections example: Frost (1972)

Subject's task: Encode 16 pictures

Instructions:

Group 1) You will need to recognize the actual object

Group 2) You will have to recall the names of all objects seen

Results:

Subjects in group 1 faster at responding to exact same drawings

Subjects in group 2 did not encode exact pictures

Conclusion: Evidence for explicit encoding strategies

Encoding of information in STM and LTM

Encoding of verbatim material vs. gist

Verbatim representation is lost very quickly (within a few seconds)

Johnson, Bransford & Solomon (1973)

John was trying to fixthe bird house. He was looking for the nail when his father came out to watch him and to help him do the work.

Subjects have false recognition on:

John was using a hammer to fix the bird house when his father came out to watch him and to help him do the work

We make automatic inferences

Does that mean we can't encode verbatim information into LTM?


Limits on Encoding

Attention: we encode only what we pay attention to:

Incomplete sensory processes: Not all information gets in through the sensory processes

Motivation; We encode what we want/need to encode

Biases in encoding based on interests/background

Improving encoding

Organization: Organize the information in a meaningful way

Ellis & Hunt (1993): Items organized (blocked) in categories produced better recall than those not in categories

Bousfield (1953), Tulving (1962) subjects will impose structure or categories on information to help with encoding.

Chunking: Chunk information into meaningful units

Give examples of memory experts. S. F., Rajan

Warmup: Practice with similar material.

Distributed practice

The Spacing Effect and Spaced vs. Massed practice


Overlearning

Overlearning increases speed of access, probability of correct retrieval

Large knowledge base: Encoding depends on having the appropriate prior knowledge

The llama's hemoglobin has a left-shifted oxygen-binding curve, which means that it can become 100 percent saturated with O2 at the low P02 values at high altitudes.

What "monetarists" do believe is that controling the behavior of M will help much to control GNP - for the reason that the changes in V will be so predictabe as to make one confident that dollar GNP will still move in the same direction as M.

Ebbinghaus found that the spacing of practice is a potent task variable, with a brief exposure providing for minimal encoding operations, but continued exposure allowing for more and deeper encoding.

Meaning: Encode information so that it ties into other parts of your knowledge.


Levels of processing (Craik and Lockhart 1972)

Levels of processing emphasizes one memory store

Conscious portion of memory called primary memory

Assumption: All incoming stimuli can be processed at different levels of analysis

The strength of a memory association is based on how "deep" it is encoded.

Maintenance vs. Elaborative rehearsal


* Maintenance rehearsal: Maintain information in primary memory


* Elaborative rehearsal: Deeper analysis of the information. Develop meaningful associations

Evidence for levels of processing

Incidental learning task

Rate the words on this feature:

Structure: Is it in capital letters?

Phomemic: Does it rhyme with bag?

Pleasantness: Is this a pleasant word?

Semantic: Does the word fit into this sentence:

John went to the store and bought a ___.

Self Reference: Does this word describe you?

Then surprise recall test.


Additional support: The generation effect (Slamecka and Graf 1978)

Generating information results in better retention than just studying it.

Subject generates rhymes to words:

TAME -----> NAME

BANK ----> RANK

FABLE -----> TABLE

....

other subject studies same list of pairs

Recall of second list of words is better for subjects who generate list than those who just studied list

Effect for math problems as well (Crutcher & Healy, 1993)

Why is there a generation effect?

Generating items results in deeper processing

Implications for studying: Better retained if you have to generate material (e.g., paraphrasing, putting in your own words)


Why do deep levels of processing promote retention?

Hyopthesis 1: Elaboration: We generate a rich elaborate encoding with many associations

Hypothesis 2: Distinctiveness: Deep processing makes that memory trace more distinctive from others.

The von Restorff effectl Items that stick out are easier to remember.

Criticisms of levels of processing

1) What do we mean by depth? Hard to measure depth

2) Some evidence that maintenance rehearsal leads to long-term retention of information (Nairne 1983)

3) Transfer-appropriate processing

Words encoded for rhymes are better recalled than words encoded semantically if a subject is asked to use them for a rhyming-related recall task.

Moral: The best type of encoding for information is to encode information in the way that you will need to retrieve it.


Retention and Remembering.

Task: Encode the following 32 digits

23157548590183725993762497088695

Why do we forget?


* Failure to store/encode


* Decay of information in memory


* Failure to retrieve information from memory

Do we have permanent memory?

Ebbinghaus' Forgetting Curve

Ebbinghause studied 1200 nonsense syllables (13/list)


* Learned first to two correct recitations


* Measured how much he retained (savings)


* Examined the number of trials/time to relearn list


* Looked at different retention intervals

Decay of recall follows a power function

Does forgetting curve hold for other types of information?


* Meaningful vs. meaningless information


How long does information stay in memory?

Permastore: a relatively permanent storage of knowledge

Bahrick (1984) Retention of unused information over long periods of time


* Memory for vocabulary from high school/college spanish courses


* 773 students over 50 years


* Measured performance as percent of original score

Linton (1975) recall of autobiographical memories.

Why? are autobiographical memories recalled better?

Retention of non-declarative memories

Declarative vs. non-declarative memory

Non-declarative memory: not accessible to consious awareness

Declarative memory: facts that can be learned, often in one trial

Not as much loss in non-declarative memories.

It can depend on number of trial/ptractice, though


Why do we forget?

Organic disorders

Disease, damage, drugs

Disuse (Decay Theory)

Thorndike (1914) law of disuse: Unless memory is attended to, it decays

New theory of Disuse(Bjork & Bjork, 1992)

Assumptions:


* No limit on how much information can be stored but... limits on ability of retrieving information


* Information becomes increasingly inaccessible unless periodically retreived (even if well learned)

Forgetting is due to new information competeing to come to mind over the old information.

Approach is biologically adaptive.

We tend to need new information, not old information.

Interference Theory

Intervening information/events interferes with previously learned information


Evidence for interference theory:

Recall of information before/after sleep (Jenkins & Dallenbach, 1924)


* Studied nonsense syllables then recalled after full 8 hour day, or 8 hour sleep

Two interpretations:


* 1) Sleep helps with consolidation of information


* 2) Events during the day interfere with recall

Other evidence for consolidation:

1) Effects of REM sleep

2) Experiment on rats and amnesia Duncan (1949)

Trained rates on using avoidance procedure.

Varied interval between training and delivery of electroconvulsive shock (ECS)

The longer time between learning and ECS, the better learned.

Types of interference: Retroactive vs. proactive interference

Retroactive: new info interferes with old info

Proactive, old info intereferes with new info

Effects of stimulus similarity on interference


Availability vs. Accessibility

There is a distinction between whether something is accessible or available

Memories in LTM are often available, but not always accessible

Evidence for accessibility/availability distinction

1) Motivated forgetting

When we want to make a memory less available

Types of motivated forgetting:

Supression: Consciously tries to forget information

Laboratory experiments: of directed forgetting (Bjork, LaBerg & Legrand)

Represession: Unconsciously lessens accessibility of information

Freud's view

Dissociative disorders

2) Cue-dependent forgetting

Memory failures can be due to not having the right cue

Example: Tulving & Pearlstone (1966)

Two groups of subjects learn list of words from categories

Furniture: chair, bed, table, sofa, ...

Fruits: banana, apple, orange, kiwi, ...

One group given category name before free recall, the other not.

Result: Improved recall if given category name

Implications: Items were probably available to both groups, but without category name, they were not accessible.

Suggests that proper cues help recall.


3) Tip of the tongue phenomenon

A person's failure to recall information that the person knows he/she knows.

What is the word designating the small boat used in the river and harbor traffic in China and Japan?

Evidence for availability but not accessibility


4) Context-dependent memory

Encoding specificity principle: Information will most likely be retreived if it is paired with the same contextual cues as when it was learned. (Tulving, 1974)

Similar to transfer- appropriate processing: The most appropriate encoding for information is to encode it in the way that you will need to retrieve it.

Evidence: Light & Carter-Sobell (1970)

"The STRAWBERRY JAM tasted great"

S's told they would have to recognize underlined phrases

Recognition test, did you see this noun before?

STRAWBERRY JAM recognized 65% of the time

TRAFFIC JAM recognized 27% of the time

Other context effects


* Abernath (1940) Subjects who study in the same classroom, do better on tests


* Schab (1900) subjects study in room with or without chocolate odor

24 hours later, subjects perform better on surprise test with matching odor


* Godden and Baddeley (1975) Underwater vs. above water context effects

State-dependent memory

Mood-dependent memoryRetrospective and Prospective Memory


Retrospective memory: ability to remember past events/information

Prospective memory: ability to plan/remember future actions


* Take a pill at 5:00


* Take your umbrella today


* Call you mom and say that you got home safely

There tends to be a zero or negative correlation between prospective and retrospective memory performance

Difference of remembering what vs. when.

Types of Prospective memory

Habitual vs. Episodic

Much of memory requires use of both prospective and retrospective memory

What factors affect prospective memory?


* Having appropriate cues and reminders


* Compliance, motivation and commitment


Click here to go back to course web page.