-Subjects studied 612 pictures for as long as they wanted

-test immediately, after 2 hours, 3days, 1 week, and 4 months 

-Results

-Immediately and after 2 hours: virtually perfect

-1 week: 87% correct

-after 3 months: still above 50%

-Small capacity (7 +/- two"chunks")

-short duration (seconds)

-what youre thinking about right now

-very large capacity (no known limit)

-very long duration (lifetime)

-you dont have to be thinking about it right now

-older concept

-emphasizes passive storage of information

-memory for a small amount of info for a few seconds

-newer concept (developed out of STM models)

-emphasizes both storage and active processing of info

-memory for small amount of info for a few seconds 

-digit span/spatial span

-delayed recognition

-backward digit/spatial span

-reading span

-hear a series of digits, immediately repeat back in order OR in reverse order

-number of digits to repeat increases on each trial until subject can no longer repeat back correctly 

Digit Span= highest number of digits subject can repeat back without error

1. Encoding: present letters

2. Retention (1-20sec)

3. Retrieval: was this letter there?

-read a series of sentences out loud, then recall the last word of each sentence in order

-reading span: highest number of words subject can repeat back without error

-evidence for the independence of two systems

Ex: System A-Task 1, System B-Task 2

-lesion A, impair task 1

-lesion B, impair task 2

-inability to recall info aquired AFTER brain injury

-cant form new LTM

-Ex: 50 first dates 

-inability to recall info aquired BEFORE brain injury

-cant remember your past 

Patient H.M.

-neurobiological evidence 

-Anterograde amnesia

-impaired LTM, normal working memory

-a SINGLE dissociation 

-Amber got an A in Psych 240

-Ben got a C in Psych 240

-Both students got an A in psych 111

-Because Ben did well is psych 111 but not psych 240 means these classes require different skills =WRONG

-normal LTM

-impaired working memory

-single dissociation 1= Patient H.M.

-single dissociation 2= Patient K.F.

-Double dissociation

-established independence of WM and LTM

-listen to 15 words then write down in any order (free recall task)

-Percent recalled: high for first word, then drops, then high for last words

-primacy=first words (LTM)

-recency= most recent words heard (WM)

-listen to 20/30 words

-count backwards for 30 sec from 100 by 3's

-Now, write the 20 words in any order

-percent recalled: primacy lower than when tested immediately, and no increase for recency words

-Listen to 20 words presented quickly, try to remember

-write down in any order

-listen to 20 words presented slowly

-write down in any order

-Delay before test: WM tasks (recency)

-Presentation rate: LTM tasks (primacy)

-two info storage buffers:

-phonological loop (stores info in speech like form)

-visuospatial sketch pad (stores visual and spatial info)

-Central executive

-computational operations on info

-allocation of attention to taks

-acts on info in the storage buffers 

-info storage site

-buffer capacity is 5-9 chunks

-without rehearsal, info decays from the buffer in 2 seconds

-like a "tape loop"

-re-circulates/refreshes/rehearses the contents of phonological buffer

-recodes non-auditory (visual) info into phonological code

-can hold 2 sec worth of info (however much you can say to yourself)

-info coded in the form of speech bases sounds

Phoneme: sound unit of language

-two kinds of lists: confusable, non confusable

-confusion occurs if words sound alike

-NO confusion for similar looking words

-NO confusion for similar meaning

-prevents formation of phonological code

-inhibits the subvocal rehearsal process

-ex: repeatedly say "the"

-With articulatory supression: lower accuracy, and acoustic confusion effect DISAPEARS 

Word length: memory span better for short words than longer words

speed: memory span better for words pronounced quickly, for people who speak quickly, and in languages where words are pronounced quickly 

-time effects due to decay

-chunking effects due to capacity of store

-time effects due to speed of rehearsal 

-give delayed response task, measure memory condition and control condition

-PET activation in delayed response minus PET activation in control task = brain areas related to phonological rehearsal 

Phonological WM (left hemisphere)

-articulatory loop: Brocas area

-phonological store: Inferior parietal lobe

Visuospatial WM (right hemisphere)

-inferior frontal

-occipital 

Frontal lobe syndrome

-distractability, difficulty concentrating

-perseveration: fail to inhibit inappropriate behavior

-problems with organization, planning

-conscious recollection

-declarative knowledge (knowing "that")

-episodic (personal experiences)

-semantic (facts)

-unconscious change in performance

-procedural knowledge (knowing "how"

-priming

-classical conditioning (the office altoid exp.)

-direct

-require subject to make explicit reference to previous study or learning

-learning is tested through recall, and recognition

-indirect

-tests do NOT make explicit reference to previous study or learning

-learning is tested through changes in performance

-relatively rare in the real world

-due to psychological, not physical harm

-retrograde (cant remember before)

-amnesia due to head injury, etc.

-anterograde (cant remember after injury)

-lives in the present only

-music can extend his period of consciousness

-anterograde amnesia and nearly complete retrograde amnesia for episodes

-no new memories

-cant remember people except for wife

-no episodic memory

-some retrograde semantic memory in tact 

-causes severe anterograde amnesia

-damage to the mammillary bodies associated with chronic alcoholism

-confabulation (making things up) is a common symptom

-Participants: Korsokoff's amnesics and NA versus normal

-Methods: experiment included 50% repeated words across 4 days

-non-repeated words: implicit

-repeated words: implicit + explicit

-Results: new words (improved the same, implicit)

old words (normal better than amnesics, implicit + explicit)

-test number of moves to reach solution

-amnesics and control performed similarly 

-(one type of implicit memory)

-a previous encounter with info facilitates later performance on the same info

-facilitation in the processing of a stimulus as a function of a recent encounter with same stimulus 

-can occur even if there is no conscious recollection of having encountered the info before 

-priming=complete more old fragments than new

-result: amnesics show normal priming, but poor recognition memory (they dont remember having seen the words)

-Korsokoffs impaired on free and cued recall tests (explicit tests)

-perform like normal subjects on word stem completion test: measures priming (implicit tests)

-subjects given word list, and word stem list

-Explicit test: say the word from the list that began with that stem

-Implicit test: say the first word that comes to mind

-explicit results: hippocampus and frontal lobe activity increased

-implicit results: posterior visual area activity decreased

-study list of words (half visually, half auditorally)

-half subjects get explicit memory test (visually presented, yes-no recognition test)

-half subjects get implicit test (word stem completion test, tests priming)

-results: modality of presentation affects implicit memory performance, but not explicit 

-Explicit test phase: visually presented, yes-no recognition test

-Result: no difference in performance for words that had been visually presented and words that had an auditory presentation

-implicit test phase: visually presented priming test. subjects had to name the word as fast as possible

-results: faster at naming the word if the word had been visually presented (same modality as test phase)--priming results when the word appears in same modality 

-words can be encoded at different depths (levels)

-physical appearance of letters

-acoustic sound of word

-semantic meaning of word

-deeper encoding=easier to remember

-half subjects do deep encoding task

-rate each word for pleasantness 

-word meaning must be processes

-half subjects do a shallow encoding

-respond with words that rhyme

-no need to process meaning

-Subjects given a memory test, half get explicit, half get implicit

-results: deep processing have better accuracy at explicit task than shallow processing

-no affect of level of processing on implicit test

-general knowledge about the world, facts

-not tied to a specific time and place (not episodic)

-organized as a hierarchy of concepts/categories (like files on a computer) 

-Goal: design a computer model that can understand language

-developed memory structure, answers general knowledge questions

-wanted the model to be able to verify sentences that require general knowledge about the world

-each "node" a concept

-nodes connected to other nodes, organized hierarchically

-features of each concept connected to each node

-properties stored at highest level ("can fly" with bird not robin)

-model searches through knowledge by "traversing" across links between concepts

-more links=more time

-model responds "true" if category membership or properties can be directly confirmed

-model responds "false" if direct conflict with a property, or category membership is mutally exclusive 

-if model is organized like human memory, then questions that require to traverse more links in memory should take longer to respond to 

-category judgements, more links=more time

"shark is animal" longer than "shark is a fish"

-property judgements longer than category

-distance effect: items separated by more links take more time to verify

-property judgements take longer to verify than category judgements

-Reverse distance effect

-typicality effect 

-revised TLC

-concepts represented as nodes

-associations (links) bw concepts vary in strength

(strength represented by distance bw nodes)

(similar concepts are close together)

-concepts NOT organized hierarchically

-links bw concepts labeled 

-nodes activated, spread to adjacent nodes

-can explain the reverse distance effect and typicality effects(typical words are closer to the concept node)

-evidence for spreading activation model

-prime activates "doctor" node,

-activation spreads to "nurse" node,

-"nurse" node already activated so fast response 

-category specific defecits of living things and non-living things

-alternative explanation: property based organization 

-semantic knowledge for diff properties stored in brain regions that process that type of info

-rep of living things based on visual prop.

-rep of non-living things based on functional prop.

-apparent category-specific deficit arise from damage to brain regions that represent either visual or functional properties 

-How stuff makes it into episodic memory

-Factors that influence encoding:

1. attention

2. processing for meaning

3. self-generation

4. spacing

5. organization

-subjects memorized lists of 15 auditorily presented words under:

-full attention (no secondary task) or divided attention (do a visually presented choice RT task)

-after a delay, memory tested using free recall

Results: less attention-->poor encoding-->poor memory

-encoding is an automatic byproduct of attending to and processing a stimulus

-the more deeply we process, the better encoded

-shallow: analysis of superficial features

-deep: analysis of stimulus meaning and connection of meaning to other knowledge in memory= elaborative processing

-Task: learn word lists

-Two independent variables: type of encoding (intentional/incedental) and depth of stimulus processing (shallow/deep)

-memory tested using free recall task

Results: Intention to memorize NOT important, depth of stimulus processing important 

-memory is better when encoding is spaced than when it is massed

-Task: learn a list of 112 words, from 4 categories

-Two groups:

1. words presented in organized trees

2. words presented in random trees

-resluts: better remember if organized 

-external context (context dependent effect)

-internal states (state-dependent effect)

-retrieval is better when the physical environment of retrieval matches that when you encoded 

-deep sea divers learn 20 auditory words

-two independent variables:

1. study context: on land/under water

2. test context: on land/under water

results: recall better when learning and testing contexts are the same 

-retrieval better when a person in the same state

-learn new words while stoned/sober

results:

-overall sober did better 

-those who studied stoned did much better when tested stoned 

-knowledge of what is involved in  particular experience/situation

-stored in semantic memory

-similar to semantic network, but more task oriented

-a type of schema

-a sequence of actions that describes an activity 

-scripts/schemas can enhance memory or distort memory

-semantic memory influences episodic memory

-washing clothes study 

-read stories like "washing clothes"

-Three conditions:

1. not shown title

2. shown title before reading

3. shown title after reading

-recall story 

-better recall when shown title before

-same as washing clothes study except instead of title, some subjects are shown a drawing that explains story 

-results: the illustration only helped when presented BEFORE the text

-means that schema must be influencing encoding

-schema is activated in semantic memory if subjects see title/picture before reading

-schema improves encoding by improving comprehension (deeper processing) and organized framework

-participants seated in office waiting for experiment, taken into diff room, then asked to write down what they saw in the office

-results: people remembered objects consistent with an "office schema"

-false memory for things that were not in office but are in an "office schema"

-prior knowledge (scripts/schemas) influence episodic memory

-memory is constructive

-semantic and episodic memory interact

-accurate, rote production of info from memory

-more likely to occur with simple materials (letters, digits)

-the active process of filling in missing elements or otherwise transforming info while remembering

-more likely to occur with material rich in meaning 

-subjects read native american folk tale and asked to recall story 

-subjects made story more coherent within their own cultural framework

-three types of errors (omit info, distort facts, add details) increased with more time between story and recall 

-subjects viewed picture and told person next to them what they saw

-results: final reports said black man instead of white man was holding the razor 

-stimuli remembered are complex sentences:

"the girl who lives next door broke the large window on the porch"

-study phase: 4 types of sentences presented (1,2,3,4 ideas per sentence)

-recognition phase: two diff types of sentences presented (old/new)

-varied the number of ideas in the test sentences (1-4)

-equal confidence for old/new sentences

-more ideas, more confidence

-memory of info is never verbatum

-summarize content and integrate details into higher-order units (constructive)

-extract general meaning and discard the details 

-people tent to remember inferences they made as if they were actually explicitly stated or as if they actually occurred 

-true for logical and pragmatic 

-inference: going beyond available evidence to form a conclusion

-logical (deductive): a conclusion that MUST follow from the available evidence

-pragmatic (inductive): a conclusion invited by the evidence but not logically demanded

-saw assertion and implication commercial

-1/2 of subjects that saw implication commercial claimed they heard the commercial make an explicit assertion (they didnt)

-the questions eyewitnesses are asked about an event can alter their memory for that event

-old info is overwritten by new info (from misleading questions)

-old info is compltely gone from LTM

-Loftus misinformation tests:

-smashed vs hit

-yield vs stop sign

-misinformation is incorporated into memory representation for event

-accurate info remains and co-exists with misinfo

-misinfo effect is due to SOURCE CONFUSION

-start with open-ended questions

-reinstate conditions (original context)

-allow witness to tell story without interruption

-use reverse order for specific questions (reduce possibility of inferences)

-sequential presentations, NOT simultaneous 

-simultaneous encourages relative comparison

-sequential reduces false ID

-not a good technique

-witnesses anxious to cooperate so generate lots of info even if its not accurate 

-Loftus misinfo effects increase under hypnosis 

-Event related potentions (ERP)

-guilty knowledge technique: only criminal has memories from crime, getaway car was a Chevy

-guilty knowledge will cause P300

-600 words studied as long as they wanted

-tested immediately and after 3 days 

-results: word recall tested immediately the same accuracy of picture recall tested 1 week later

-word recall tested at 3 days same as picture recall tested at 4months 

-estimated if shown 1,000,000 picutres, could remember 730,000 of them 2 days later 

-verbal memory much more limited 

-Experiment 1: draw each side of penny from memory (only 1 accurately recalled)

-Experiment 2: recognition, only 50% who got accurate drawing correctly recognized it 

-for unimportant and unattended details 

-when stimuli lack meaning or relavence

-when alternatives (incorrect choices) are similar 

-Need to attend to details, meaningfulness of details, and distinctive alternative 

-visual code has more info than verbal code, so visual memory is better than verbal memory 

-more details result in more connections to other things in memory 

-4 stimulus conditions

-memory test: old/new recognition test

-immediately and 7 weeks after 

-results:

-memory for all 3 picture types better than memory for verbal descriptions

-no difference between 3 visual conditions (suggests amount of detail isnt critical factor)

-we use both verbal and nonverbal codes to represent visual info 

-visual material usually represented with both visual and verbal code

-verbal only have verbal code 

-concrete words (verbal and nonverbal code) are remembered better than abstract words (only verbal code)

-method of loci: associating visual images with verbal info 

-bad for unattended details because we dont create a verbal code for those, lose dual-code advantage

-bad for when incorrect alternative are similar

-visual code doesnt help distinguish them from correct alternative

-have to rely on verbal code 

-3 conditions in recognition test:

-lure item is novel

-lure item is exemplar from the same category

-lure item is the same as the picture you saw but in a different "state"

-objects remembered nearly as well in exemplar and state conditions as in novel condition 

-verbal code must carry deep conceptual meaning, cannot just be a word

-internal representation of our physical environment that store the spatial features of the external world

-rats in maze: rat given no reinforcement for first 10 days, but reinforced after 11th day

-wondered aimlessly, but when given reinforcement learned the maze right away

-must have used a cognitive map

-weaknesses of cognitive maps

-mental shortcuts that can distort our cognitive maps

-right angle bias

-symmetry 

-rotation

-alignment

-relative position

-subjective clusters 

-generation

-maintenance

-transformation

-inspection 

-visual imagery is seeing in the absence of the appropriate sensory input

-"perception" of remembered info

-retrieval of info about object properties from memory

-will this container fit in the fridge? will the car fit in that spot?

-all mental representations are in the form of abstract propositions

-proposition=deep meaning

-we recreate verbal and visual codes FROM propositions (words and images are epiphenomena)

-one way to represent propositions

-describes logical relationships between entities 

-basic form: 

[relationship] [subject, object]

-ex: On (Cat, Table)

-imagery isnt just like perception:

-hard to reinterpret ambiguous figures using imagery

-images dont have metric properties

-hard to determine part-whole relationships using imagery 

if perception and imagery tasks interfere then they must require the same mental system 

-we represent and use mental images in ways that are analagous to our representations and uses of perceptions

-image scanning and scaling experiments

-mental rotation experiments 

-study pictures, memorize them

-ask to focus on one end of the picture

-Results: RT to find far object parts longer than RT to find near

-suggests "scanning" the image

-demonstration of functional equivalence: we scan mental images like we scan real images 

-experimenter's expectancies about how subjects are going to perform create an implicit demand for subjects to perform as expected 

But...

-if you tell experiments to expect something diff, still get original outcome

-if you tell subjects to expect something diff, still get original outcome 

-subjects faster to answer questions about visual properties in the NEAR condition than in FAR condition

-demonstration of functional equivalence: details on near objects easier to see than details on far objects

-judge if 2 3-D objects are the same or diff

-3 conditions: match 2-D rotation, match 3-D rotation, or no match

-demonstration of functional equivalence:

-takes longer to rotate objects at larger angles (doesnt matter if in plane or in depth) 

-both for real objects and mental images of objects

-interference effects

-size judgements

- Functional equiv. (scanning, zooming, rotation)

1. basic code is propositional (for long term storage)

2. propositional code used to create depictive image

3. depictive image can be scanned, zoomed, etc

-had subjects do a visual imagery task and a visual perception

-brain activity in visual (occipital) cortex for BOTH perception and imagery (share common mechanism)

-more processing required to create an image than to perceive it

-fMRI study: perceive or imagine faces of buildings

-Results: ventral temporal lobe activation for perception and imagery for both types of objects