Week 2 Lecture 1 Short Term Memory example essay topic

Cognitive Psychology Week 1 - Lecture 1 Cognition - matching the world to internal representations - language and word recognition - pattern recognition - visual imagery - memory we will focus on memory - problems - when cognitive processes go wrong brain damage dyslexia amnesia - viral infections, car accidents, severe alcoholism, old age (alzeihemers) - they don't lose knowledge of the world, but once it has gone beyond immediate awareness you cant recognize / remember what has actually happened to you. Memory is relevant to clinical psychology False memories - how they occur and how to distinguish b / w false and true beliefs Transfer appropriate processing - if you match retrieval process in memory tests to the way in which you learn. Cognition Process Conscious and subconscious processes. Conscious - reminders we give ourselves to remember events. Unconscious - no conscious insight of how processes work, cant study by introspection - no conscious awareness. How to study cognition. experimental psychology, accuracy or reaction time in normal subjects...

At the mind level - brain function - neuroscience done on animals - but not much done on cognition... Studying brains in action in humans 'a functional brain imaging - what aspects of brain activated in certain tasks. Cognitive neuropsychology - look at those with brain damage in specific areas - localized brain damage which affects cognitive processes in one area, but not others i.e. long or short term memory - is their overlap in processes?? Computer processing - Wont cover Week 1 - Lecture 2 MEMORY 1.

Definition of memory - Result of learning i.e. modification of behaviour and / or knowledge through experience. Complicated by 2, 3, 4 2. Stage of memory - stages of processing. Encoding (acquisition) - study phase (phases of experiments on memory). Storage (retention) - distractor phase.

Retrieval - test phase Encoding - learn material - at a party and processed info. Not deliberate study Storage - before you try and access memory it must be retained in this intervening period Test phase - trying to remember what happened Overall memory performance depends on all 3 stages and interactions b / w them. How could memory fail?? - could happen independently at any of these 3 stages - fail to encode - not stored - test - retrieval - couldn't find right cue - it is much more complicated than this 3. Examples of memory -many different things are referred to as memory.

Recall a list of words. Recognize a familiar face (retrieving). Remembering your 4th bday. Remember a phone number long enough to dial it.

Remember the name of an object (word finding difficulty). Skills. Lean to ride a bike / learn to read. Remember what you have to do tomorrow (prospective memory) remember you have to buy bread on way home from work - wont cover this type of memory 4. forms of memory (retrospective) - already happened - these examples are al related but not the same cognitive phenomena. They are functionally independent but not completely independent (some) at mind level.

Systems of memory also independent at brain level. Schacter and Tulving (1994) - memory systems a) short term / long term memory - Peterson and Peterson; Baddeley 1960's and 70's - primarily based on duration - info retained transiently or permanently tasks: repeat immediately (STM) memory span (STM) recall after 2 week delay (LTM) b) episodic / semantic /procedural - Squire; Tulving 70's and 80's - primarily based on type of info stored - personal events (episodes) episodic vs. general knowledge (semantic) vs. skills (procedural) tasks: recall studied paragraph or film identify famous faces reading inverted text c) Explicit / implicit memory - Schacter, Jacoby 80's and 90's - based on conscious / non conscious nature of retrieval - information recollected: "explicit" vs. expressed on behavioral measures "implicit" tasks: recall, cued recall, recognition priming on stem completion (first word that comes to mind) CAL Priming as decrease in RT's in lexical decision or naming 'a present something on screen either real or nonsense words and measure speed taken to make decision whether real or nonsense. If it comes up again after a delay - we are quicker (this is an index of implicit memory - even though you may not remember whether that word was shown twice or not) independent of conscious recollection. EXPLICIT vs. IMPLICIT Not based on type of info d) overlap in distinctions. different distinctions within memory are not mutually exclusive. e.g. You can have explicit retrieval of info from semantic memory etc. semantic - generalized piece of info. priming - picture of prince Charles - then picture of princess Diana - quicker to say she is a royal over a politician - primed by prince Charles in previous trial (implicit). prosopagnosia - inability to recognize faces - brain damage, localized disability or could be developmental... If gave a picture of prince Charles - wouldn't be able to give name. don't have explicit consciousness but do have unconscious ability (implicit). Many combinations of these distinctions.

ST, explicit - conscious measure. ST, implicit - priming effect. We cant have conscious recollection of procedural memory - ie. Riding bike 5. forms vs. psychological systems - different storage?? - Or different methods of access to common storage system?? - Neurological basis of distinctions??

Yes there is evidence of neuro basis to these distinctions but info being stored (explicit / implicit ) is not in different forms - single storage system but 2 brain systems which allow access either directly (conscious) or indirectly (unconscious) 6. This course - what is memory / how do we remember complex questions - must define type of memory you are interested in. - consider each form independently - NOW - we will focus on explicit memory and memory for episodes (transient) and divide this up into - transient (ST) and enduring (LT) Week 2 - Lecture 1 Short term memory / working memory Evidence for short term / long term distinction 1. limited memory span - present series of digit to subject - increase by one each times - ask them to repeat back to you - "span" = no of items that gives perfect recall (in correct order) 50% of the time - either half of subjects tested or half of trials on individual subjects - where does subject fail? At what point do they fail DID A LITTLE MEMORY SPAN TASK Work out last one you get correct -the test doesn't continue once you have failed There becomes a point where you cannot recall Drop out b / w 6-8 digits Capacity of STM seems to be around 6-8 - Miller (1956): 7 +/- 2 - reason for calling it a magic number was that this same number came out as memory span for almost all subjects for a whole range of stimulus classes - binary, decimal, letters of alphabet etc. 1 syllable English words.

- Keeping them verbally active - Capacity of short term memory was 7 "things" - Span of 5-9 items with binary digits, decimal digits, letters of alphabet or English words. - Span based on number of "chunks" e.g. 1945 vs. 1386 - history effect - might use strategy of 1945 end of war and you remember that as a meaningful chunk rather than disconnected digits - Span for letters is o 6 if selected at random o 9 Consonant-Vowel-Consanteant syllables - if organized in this way, they are pronounceable and can use this as a strategy to remember o 50 in meaningful sentence - can repeat back a large number of letters if it is in a meaningful sentence... - i.e. perceptual and recognition systems identify familiar units (chunks) and STM holds 7 +/- 2 of these. Its about chunks - earlier perceptual recognition, semantic processing affects ability of span. We do have a limited memory span - verbal rehearsal, before they disappear. Traces are transient, different from long term processes - where we have encoding that is not difficult, it is just done. What is a thing?

A letter is one letter, but words have many letters - CHUNKS 2. Recency Effect - - lists of unrelated words - Postman and Phillips (1965), Glazer and Cun itz (1966/). given a random list of words, asked to remember a lot of those (about 30). pushed beyond memory span - impossible to repeat all back, cant keep them in STM. immediately told to report words back in any order that you like. plotting accuracy - serial position curve. on x axis - serial position - where original item appeared. words presented first remembered well. words in middle remembered pretty badly (3-27) 15% recalled correctly. words presented at the end are remembered increasingly well - last item almost ceiling levels (80%). Recency effect - those presented last - retrieval from (working memory) short term memory. People spit out last words first, they are active in working memory, those other words have been pushed out, they cant remain active, others replace them... Primacy effect - those presented first -why are first ones learned better?? Primacy effect probably due to proactive interference - as you have to learn more in a list, it becomes harder to learn each successive one as the previous one interferes- this is a property of long term memory, not important as a distinction b / w STM and LTM.

Most retrieval for words presented previously is from LTM, working memory is better for recall than long term memory. Limited short term memory store. i.e. limited STM sore which is Dumped" first in recall of list; will most often contain last few words - order of recall: last first. people write down the STM memory words first... This is reasonable good evidence from STM component separate from LTM Double Dissociation in SPC (serial position curve) components (a) brief filled delay -. manipulation introduced - give list of words, then told 739 and asked to count back 3 out loud - this is used to stop us keeping those last few words in working memory... This will get rid of recency effect, but if only a relatively short delay, this shouldn't affect performance on long term memory task. The lower two graphs have done this... Do this for 15 seconds or 30 seconds - recency effect disappears but doesn't get rid of long term effect, if subject not given another task, then you do not lose recency effect - we are able to keep them active by rehearsing them.

The 30 second delay is the same as for 15... when you aid in filled delay, performance doesn't drop to zero - it is at level of items retained in short term memory, memory accuracy is a sum of two components those recalled from LTM, and items recalled from STM. By using a manipulation you can alter short term memory but leave LTM the same. no recency effect, LT value delay (b) Glazier (1972). showed recency unaffected by word frequency, recency was unaffected, but influenced earlier part of serial position curve. word frequency, including low frequency words, presentation rate effect long term memory, age of participants effects long term memory but not recency part of curve, divided attention (doing 2 tasks at once) has a big effect on ltm, but has little effect on stm... There are functionally distinct forms. At the level of the mind, they are distinct. There is a double association that you can demonstrate. What about at brain level?

YES!! Looked at through brain damage to different areas of brain. Areas which affect long term memory and areas that affect only short term memory Lecture 2, Week 2 3. Neuropsychology Amnesia e.g. HM (Milner, 1966) - patient called HM - late 20's, 1956, given brain surgery for epilepsy to stop epilepsy, it worked helping epilepsy, but couldn't retain anything in LTM - trouble with Long term memory - could remember pre trauma events and knowledge - never heard of president Kennedy - cant retain any info in long term memory - he can implicitly, but not explicitly - he has been tested over and over, never recognizes testers - cant learn new material - normal span - perfectly intact working memory - one on one interaction OK sentence to sentence - can carry on normal conversation as long as his attention isn't distracted. Must be about neutral things, as soon as he is distracted, you will have to reintroduce self if you go out of room for 5 mins. - cant transfer info to long term memory Baddeley and Warrington (1970) Serial position curves for amnesiacs - normal recency effect, poor LTM 'e have STM ok but cant add to LTM 'e copy of this on handout 'e learn 30 words - show normal memory for those words at end of list 'e in HM shows high recency effect but cannot remember those items presented more than 4 trials ago (almost zero) STM damaged patients - Shall ice and Warrington (1970) - patient KF - Digit span of only 2-3 digits - LT learning normal - asked to learn list of 30 words, and recall 30 mins later, this was normal - Serial position curve shows opposite pattern to amnesia - early part normal but severely reduced recency effect - looked more like filled delay - Have LTM ok but damaged STM Together - Double Dissociation at brain level This adds a lot to evidence for physically and functionally distinct systems - STM and LTM 2 implications 1. methodological - a ST task testing over short delays will be tapping both short term and long term memory - sum of STM and LTM to give total performance (serial position curves) - small contribution from LTM - need to distinguish tasks from memory forms tasks are assumed to tap - not a one to one correspondence, it is easier to get a solely LTM memory task, but difficult for STM 2.

Theoretical - most active info is in STM for processing - we are paying attention to these things the most, they come from current perception but may also come from LTM - this info can be retrieved from LTM as well as from current perception - i.e. STM = "working memory" Limited Duration or capacity?? HOW DOES FORGETTING FROM STM OCCUR?? - trace decay (naturally), interference (next item pushing out previous item) or both?? i.e. 1. limit on time because need to re-rehearse item before it decays - we need to re rehearse before it decays - decays before you have time to get back to rehearse or 2. limit on number of slots to be filled: each new item pushes out old one?? - storage spaces are limited, can think of 7 items at once, once the 8th item is added, one has to fall out of slot, Or a combination or both? Previously said "limit" but didn't distinguish possible causes. ANSWER: Time limit / Trace decay - not a fixed number of slots 7+/- 2 but - exact span depends on articulation rate - Baddeley, Thomson and Buchanan (1975) o Span task - 5 items 1-5 syllables long - wit, sum, harm, bag, top vs. university, opportunity, aluminum, constitutional, auditorium o Recalled decreased as number of syllables increased o Also measured reading speed for lists o For all lists got same answer: no recalled / reading rate = 2 seconds of words o if it takes you longer to say a word out loud, it will also take longer for subvocal articulation. o Number of words you can accurately recall corresponds to number of words you could sub vocally rehearse in 2 seconds o Span is the number of words than can be rehearsed in 2 seconds o Is it duration not number of syllables?? YES - Baddeley Must test in native language Conclusion: limit on short term memory is not slots but an indirect result of trace decay of each item you are trying to remember - they begin to decay unless you can rehearse. We get capacity limitation (interference) because of duration limitation.

Is rehearsal necessarily phonological (hear) - in normals for verbal stimulus, YES, but can rehearse visual info (mental rotation on text) What if born deaf? Don't have the phonological system, don't know what words sound like Kirkham (1993) - PHD thesis - deaf native signers Looked at memory span related to how long it took you to sign the word in sign language - not articulation rate of equivalent spoken word Span was normal, but how long it took to rehearse words in sign language. Normal system - but all about how long it took to sign words. Signers internally rehearse without phonology. Suggests a number of potential "rehearsal loops" of limited duration / capacity. Different loops deal with different types of information Why do we have a short term memory system??

Week 3, Lecture 1 What is STM not for?? Two views: 1. Atkinson and shiffrin (1968) idea - WRONG - in order to remember permanently, must go through STS to get into LTS - more rehearsal = more likely to transfer (better LTM), need to actively rehearse to transfer into long term memory - functional role of short term memory was very dull, only used as a portal to long term memory, didn't have its own function - underlying assumption - saving info to long term memory is very difficult - model of r ship b / w short term and long term memory - key idea 'e functional role only to keep (some) info around long enough to be saved permanently in LTM. - "modal" model - also often called modal model - "rise and fall of STM" = rise and fall of this model, not of STM ATKINSON AND SHRIFFIN MODEL PICTURE - sensory register - very brief - 100's of milliseconds WHY IS THIS MODEL WRONG?? - Don't have to go through STM to get to LTM (neuropsychology -patient KF, digit span of 2-3, but normal long term memory, can cause brain damage that knocks out STM but doesn't affect LTM, this wouldn't occur if the A&S model was correct) - We can keep far more info in LTM that we can actively rehearse in STM - Simple rehearsal = poor LTM, it helps to some extent, but not really. - LTM improved more by processing for meaning than for rote rehearsal - Natural learning is learning for meaning - e.g. incidental learning, having a coffee with friend and remembering that conversation - Intentional learning is trying to learn. What is STM actually for??

NOT for strengthening trace via rehearsal so that info can reach LTM 1. Working memory - limited capacity system - incoming stimuli and info from LTM meet there - she finally left that jerk - current info with info from LTM - info from whatever source is in WM is currently most active current stimuli 'a sensory perception processing 'e WM 'e LTM neural mechanisms take a little while to consolidate LTM don't need to go through STM - can retrieve info from LTM and rehearse it and make it active - arrow back to LTM - FUNCTION - "working" in that processes info as well as holding it a) storage and computation - thinking, reasoning about material b) subcomponents in WM - a collection of limited capacity "stores" this makes it extremely useful. - Need to keep stuff active long enough and to integrate it with LTM a) storage and computation - if WM does both, does each function interfere with each other?? Posner and Rossman (1965) computation with storage or recall of a set of digits - gave ss 8 digits to learn: 8 3 4 4 2 5 9 1 - after various delays recall first 3 digits - we are trying to rehearse these to store them and keep them active - in the delay you are asked to do mini reasoning tasks some simple some difficult 1. recording - write down last two digits you heard over and over - easy 2. add - sum the last two and write down over and over - not just storing but processing the sum of 9 + 1 - slightly more difficult 3. backward - count by 3's from last pair 4. classification - most difficult - classify them as grater than 50 or less than 50 - then recall first three digits at delays, ss presumably rehearsing theses - the harder the computation task, the poorer the recall. Baddeley (1986) showed the opposite of the above o Storage interferes with reasoning o Does performance on reasoning task gets worse as you have to rehearse more digits in working memory Ss rehearse digits out loud: on graph referred to as "concurrent load" - how many digits told to rehearse - 61616161 or 6734673467346734 - REASONING TASK - statements True / False e.g. A follows B - BA (True) or B is not preceded by A - AB (False) - measured how accurately you could do it and also your reaction time - reasoning time increased with size of concurrent digit load (BUT not catastrophic - why?

Even with 8 digits you were still able to do it, it just took a little longer. Take home message: storage interferes with computation and vice versa - evidence for working memory - not just storage, it is actively working on it Increasing digits didn't increase error rate, it is still in front of you, you can go back and check It was the time taken that was affected by the concurrent digit load b) subcomponents in WM interference is not always found b / w storage and reasoning e.g. Halford, Bain and Mayberry (1983) - concurrent load only weakly affects performance on algebra problems - different type of reasoning task didn't have much of an affect. - given a load of digits (b / w 2, 4, 6) - then had to solve 3 x + 6 = 9 in head, whilst recalling digits explanation for this: Anderson, Reder and Lebiere (1995) - - for many people algebra problems can be imagined being done visually - maybe one occurring in verbal rehearsal loop and one in visuospatial sketch pad, and because using different components of the system, they don't interfere. Doing algebra on visual blackboard in mind - added condition where substitution required - so there couldn't be complete independence of processing ax + b = 9 - need to work out what x equals and told that a = 1st digit of concurrent load and b = 2nd digit of concurrent load once you required this substitution - the amount of storage and the performance on reasoning task did interfere with each other again latency - time taken to solve algebra problem - without substitution - no effect of concurrent load - with substitution - got large effect - get interference b / w storage and computation only to extent that tasks forced to use same processing system - weak null effects of interference found when 2 tasks can be done by different WM sub systems BADDELEY MODEL. Verbal rehearsal loop - phonological loop - verbal rehearsal of concurrent load. Visio-spatial sketchpad - used for algebra problems, when no substitution. Central executive - can transfer info from b / w these loops, this is more conscious - this is a vague notion.

Read text about this model Week 3, Lecture 2 What is STM for (CONT) 2. integrating info spread over time e.g. Language processing retaining info long enough to - integrate words into sentences - integrate key sentence ideas in text a) Use in single sentence processing - cognitive neuropsychology approach - patient TB (Baddeley and Wilson, 1988) - has STM memory problem, arose from epilepsy - very smart man - professional mathematician, memory problem after epilepsy - some LTM problems, but learning list of words was normal - very severe STM damage - digit span 1-2 - problems in real word comprehension - if Baddeley correct this STM problem should lead to problems in language processing - doesn't have trouble with meaning of single words - ok on short simple sentences - but in real world comprehension he has a bit of difficulty, he misunderstands what they have meant. - when tested experimentally - if due to problems in STM, his understanding of sentences should be fine, if they are very short and simple syntactically, so no difficult in integrating simple true or false decisions - he was quite good at this "bishops are made in factories"slippers are sold in pairs" this is greater than his digit span - so maybe he couldn't do, but when processing sentences you make an ongoing consolidation process - "bishops are made" is one concept, this is why he can do it. When it became more difficult: so poor he became distressed with verbose sentences: "it is commonly believed and with justification that slippers belong to the category of objects that are bought in pairs - couldn't tell us whether this was true or false these were the two extremes but we was tested over sentences of 1-10 words comprehension fell from 100% 'e 20% with 10 words - within sentences of a particular length, they changed the grammatically difficulty of the syntax. He performed worse on those sentences requiring higher memory load "The boy the dog chases is big" - the boy is big - but these arrived at different times, need to integrate after a time delay 'e STM used to help with long and / or syntactically complex sentences. - we only looked at this in terms of patients, but this is the same for normals. - good piece of writing - simple, well structured sentences, well integrated b) used in text comprehension - Dan men and Carpenter (1980) - measured "reading span" - measures capacity in working memory. The max no of sentence-final words that the subject can recall. - read sentences for meaning as well as for knowing they must remember the sentence final word they know they will be tested on comprehension - memory tests - given a number of sentences, this is more difficult than digit span. - some are very good - it is good at separating people. - looked at effect of individual differences in "reading span" - some people in normal range are better or worse than others. (NB realistic measure of WM ability) - ss then asked to read paragraph of text with referent (WAYNE), read some intervening sentences (GUFF IN THE MIDDLE - I like girls blah blah blah), then a pronoun - the task was to say who / what the pronoun referred to - he / she /it, who or what was the referent - see handout - manipulation was to vary the number of intervening sentences - general prediction - if WM involved in both - then performance will get worse as person has shorter reading span, and more intervening sentences. If good reading span can cope with more intervening sentences than those with poor reading span. - ss with higher spans showed best comprehension, with least fall off with more intervening sentences - do these individual differences in reading span go on to correlate with peoples ability to understand text or say a conversation - yes this is what they found 'a YES, we use WM in text comprehension as well as the reading of single sentences out loud as often they are quite long.

We use it for many other functions Conclusion - What is STM for?? O working memory storage and computation - why? For interpretation of stimuli spread out over time e.g. info that doesn't arrive all at once. (e.g. Langauge; reflecting on past events / thoughts - make it active again in WM) O Where integrating of distinct bits of info required O The "hub of cognition" - extremely important - but we have only talked about conscious cognition - one view is that working memory is consciousness long term memory and short term memory - processes, components and functions may be different, as they are independent. Week 4, Lecture 1 - 13/8 Long term explicit memory - two general approaches 1. how much is remembered versus forgotten (quantity when correct) - what factors effect how to remember 2. what is remembered (true?

False? ; types of errors) - its more than this, often we remember something that is incorrect, false memories 1. How much is remembered? Observable memory performance is influenced by encoding AND storage AND retrieval Encoding (aka study, learning) - how can we help ourselves remember info Practice - does practicing more help? At encoding stage we are holding storage and retrieval constant experiment: - recall of paired-associates [words that don't naturally go together, but have been paired in context of experiment] (e.g. Dog bike) given the first word as a cue, and then asked to recall, bike, when given the word dog. Results of this as a function of amt of practice (power function) - Anderson (1981) - as practice trials increase less errors are made. - fewer forgetting errors - if you know material better, you will be able to produce it more rapidly - paired-associates as a function of spacing of practice - Keppel (1964) - 8 times one day or twice per day for 4 days: results: better with distributed practice than massed. - see graph - interesting results: also varied the retention interval - straight away up to a week later.

As you increase delay, memory get worse, generally this is what they found. If you tested immediately after, it didn't matter if you learnt it either massed or distributed, in fact massed practice was best immediately. 1 day delay, massed practice plummeted. After 7 days, could barely recall anything. Distributed practice stayed at very high levels of remembering across the week. - theoretical interpretation: practice makes stored trace stronger in LTM, perhaps massed practice, doesn't make the practice sessions distinct. Divided attention If you do two tasks in the study phase, learning ability is detrimentally affected.

Single task - learn a list of words, in study phase, they are given a distractor task or dual task paradigm Murdock (1965) - primary task: learn spoken word list. Reason being - secondary task was a visually presented task cant do two visual tasks. Simultaneously they were asked to sort cards Varied attention demands in the card sorting procedure in easy attention demand just turn cards over, or more difficult attention demands sort cards into 2, 4, 8 categories: say divide into living non living things, 4 categories etc. - as you took more attention away from primary task being able to remember was impaired significantly - memory worse with divided attention. Levels of processing / elaboration - all about semantic processing and active semantic processing - Lab for LOP (level of processing) (Crack and Lockhart, 1972) - refers to whether you process the info for meaning or not for meaning (ie, for some other reason) learning by rote - is not learning for meaning.

Learning the alphabet - is learning for rote. Times tables also learnt by rote, rather than meaning non-semantic - does dog rhyme with boat? Semantic - is dog linked to puppy? - process for meaning - LOE (level of elaboration) - everything being studied in the study phase is for meaning, but how deeply is it being studied for meaning?? e.g. Borrow and Bower (1969) dog bike -- -- WORST vs. the dog chased the bike -- -- BETTER vs. make up your own sentence linking dog with bike -- -- BEST i.e. a generation effect - sentence constructed by you, rather than by experimenter, your generation of sentence helps you remember. 'e if you provide meaning yourself, you will remember better why is this?? Storage and retrieval phase are held constant - Anderson and Reder (1979) idea of LOP / LOE: increase memory because trace left by meaningful encoding provide ss with more retrieval routes. - an active retrieval phase, the more richly it is encoded, the more hooks you have to come back to. i.e. an effect of encoding because certain forms of encoding make retrieval easier; not because they make the stored trace stronger. You are getting a richer trace, which is easier to retrieve later on. Organization of material - Bower, Clark, Les gold and Winze nz (1969) wanted to investigate how memory was affected by type of semantic structure you provided to subjects in order to help them learn info in the first place.

Tested memory under 2 circumstances: : - hierarchal semantic structure to help aid memory - recalled 73/112 words or - random order - recalled only 21/112 for examples minerals metals and stones common metals rare metals etc precious stones rare stones etc can use structure to help you when semantic structure is provided in study phase, they recalled 73/112 words remembering structure provides us with cues to remember specific material. Intention to learn (counterintuitive result! !) no effect of intention per se. - e.g. hyde and Jenkins (1973) - level of processing study - ss rated words for pleasantness - high LOP or - whether they had e's or g's in the world - low LOP - half ss in each condition warned of a memory test (intentional learning) half not (incidental learning) - effect of LOP but not intention 'a apparent effect of intention arises because intention makes us more likely to use strategies above. O does it matter if you try to learn the material or not? O No effect of intention per se. O Intention means you use other strategies to try and remember. Week 4, Lecture 2 14/8 Conclusions from encoding Practical: all else being equal, study should be with full attention, (self) elaborated and distributed over time.

Theoretical: likely that many effects of encoding are not on trace strength, but on ease of retrieval (i.e. content of retrieval) Retention (aka storage) - not a great deal we can do in the storage phase Forgetting function - the longer the delay, the more likely you are to forget, but what is the shape of this forgetting function? - Krueger (1929) paired associate recall. - performance fell very rapidly after first week or so, after this it was slower decay - power function - shape of curve. Interval in days - power function - Squire (1989) probability of recognizing name of a tv show as a function of time (no of years since its cancellation) - given real tv show names or name of false tv show that have never existed, had to distinguish b / w the two - found a power function in shape, rapid decay after first few years of delay, after 6 years only recognize 60% and after this it really evened out. 50% is chance, because you can guess if a real show or not, so still better than chance after 15 years. Higher level of processing in this experiment. Seen the show etc etc. interval in years - power function - Linton (1982) - studied rates of forgetting everyday events.

Interval in years - linear function - everyday she wrote down a particular event that occurred in her life that day - dramatic or dull depending on the day, but something particular to that day - after a period of 6 years she went back and got someone to pull out cards from those years and then asked them to invent events, and she had to say what had occurred and hadn't occurred. After first 6 years fairly linear decay pattern. 50% is chance. There is hardly a surprise - the longer the delay, the worse memory performance gets.

What if you learn material more strongly in the first place? Do you lost it more slowly? With the same amt of study test delay Degree of learning and forgetting rate Slam ecka and McElree (1983) - extra trial of practice on sentences. Delays up to 5 days Got subjects to learn sentences for later memory test Got either 3 study trials or 4 study trials - better at remembering if you have 4 trials - did you show slower decay if you had learnt it 4 times versus 3 times? - no over the 5 days there wasn't a difference for 3 trials versus 4 trials, two lines are parallel, even though you learnt material better, the rate of decay is exactly the same for 4 trials and 3 trials, losing info at same rate!! Witted and Ebbe son (1991) - looked at memory for single words - to increase strength of trace - gave you longer or shorter exposure time to words, only go through list once, but see word for either 1 sec or 5 sec duration. Deals up to 40 seconds - still lose info at same rate, you are better at recalling for 5 sec study, but you are losing information at the same rate, when you take logs of number recalled.

The lines are parallel. But this is just rote repetition - do you get more resistance to forgetting with semantic / elaborated study?? Interference We not only have delay but we have interference b / w study and test phase. - retroactive - interference of new learning on old - there is an interference effect - retro - in the past - learned a piece of info - old - then you are presented with interfering or new info which may interfere with old. Having learnt a second thing interferes with the retrieval of the first paired associates see graph A-B-A-D A-B-C-B A-B -- rest Much worse worse Best - general finding with paired-associate learning: interference worst when given potential confusion - interference with old knowledge e.g. Bradshaw and Anderson (1982) - learnt old fact about Mozart (tested) - first thing you learnt about mozart 3 conditions 1. by itself - no interference (control) 2. with 2 conceptually unrelated facts to his trip to Paris - (worse) -makes memory worse 3. with 2 conceptually related facts to his trip to Paris (better) - makes memory better, even better than the control condition 'e new material can interfere or enhance - depending on whether or not it can be integrated with the old Conclusions from retention All else being equal, performance better if delay shorter and intervening material either very distinct from target or able to be integrated with it.

Practical implications - not a lot we can do about this. Week 5, Lecture 1 - 20/8 Retrieval - type of test Standard tests of memory in experimental setting O Recognition cued recall free recall (see text) O Free recall - not given any cues O Cued recall - given clues to help remember O Recognition - 2 different ways Old / new recognition - study 30 items - in test phase presented with 60 items, half old, half new and need to say if old or new Two alternative forced choice - study 30 items - presented in pairs one in pair will be old one new, forced choice. The more cues given, the more assistance, the more likely you will be to be able to retrieve info. Nature of distractors - recognition tests O In recognition, get worse as distractors are more similar to (less distinct from) the target.

O In recognition memory - as distractor items are more similar to target, memory gets worse. O e.g. Learn list of words - all man made objects - distractors were all animals. This would be easier - as you would know that all in study phase were not animals - how well you can reject items is also important. O Is the same true for phonetic memory?

Not really - long term memory - is more hinge nt on meaning. We remember meaning more than form in explicit long term memory. O Similarity in meaning presents problems rather than similarity in form O Line ups - examples - recognition memory test, asked to identify perpetrator. - this is notoriously inaccurate, better if you don't know how many distractors there will be. Pressure to choose one... one coming up at a time, is better. Choosing one perpetrator is easier or more difficult depending on the nature of the distractors O Cross race deficit - better at remembering faces of people of your own race. White Americans definitely show deficit for recognition of black Americans.

O Classic example: person reported they had been attacked by black man in Britain - not enough black people for lineup - they ended up presenting 1 black man (who they suspected) and 5 white policeman with black painted faces. O Important in a practical sense also!! Encoding specificity (tulving, 1972) Extent to what happens in retrieval phase matches what happens in the study phase If it matches - then memory is better than if it mismatches - match in external cues b / w study and test improves memory, if they mismatch this reduces memory - external cue = associated word - something paired with target word - something in outside world study phase: trying to remember CALENDAR, but it is paired with mother. Test phase: recall word that started with CAL Tested under 2 situations, with different cues: Mother - CAL vs. Purple-CAL Memory is better when the cue is repeated b / w study and test phase. - external cue = environmental context: Golden and Baddeley (1975) - more practical divers underwater or on land, at study test. Best memory when either studied on land and tested on land, or learned underwater and tested underwater Memory much poorer when there was a mismatch - land / water - water / land Mckone and French - not quite as extreme as the diving example, if you don't manipulate it as strongly, the effects aren't as great. Looked at environmental context effects either outside or inside.

Chifley library - inside Hancock - outside Same pattern for outside vs. inside on ANU campus - told people that she was looking at effects of mild exercise on memory and also poor budget, and honors students weren't able to have access to labs see handout - cued recall test if study inside - memory better if tested inside if study outside - memory better if tested outside what environment you were in when you learned information, and when you need to retrieve info improves memory if conducted in the same place. Transfer appropriate processing - referring to match b / w retrieval and encoding enhances memory - but focus on internal processing done by subject, rather than the external environment - emphasis on processing!! - not match in external cues per se that enhances memory, but these are effective because produce match in type of internal processing done by the subject. Emotions physiological state etc. mood dependent memory e.g. people recall more negative events from their life when they are depressed than when they are not. - same for drunk vs. sober. - e.g. our lab - Semantic -S Rhyming -S but - S-R = R-R manipulated type of processing person did rather than external cues see handout memory is better in same same than different levels of processing and study test match - transfer appropriate processing semantic study phase advantage is reduced if the retrieval test is a rhyming cue. Not exactly the same cue at study and test Dog: log, frog - not repaired with log in test phase - repaired with frog Dog: puppy, hound External cues were different, but the same type of processing was induced b / w study and test - study for meaning or for phonetics Instead induces same types of processing - semantic or perceptual Is encoding specificity really attributable to transfer appropriate processing (internal state of processing) - bjork and Richardson-Klavehn (1989) perhaps ss mentally recreate original context (e.g. imagining yourself underwater?? conscious - no evidence that this doesn't occur, but... - may help but Mckone and French (2001) found 0/28 ss used this strategy spontaneously - this didn't occur consciously - this is not an explanation for why context effects occur, - it may help. Week 5, Lecture 2 - 21/8 Conclusions about retrieval - all else being equal try to generate a cue that access distinctive nature of event and match test processing to study processing - information may be in there even if you can't retrieve it in a given test take home message: study test match enhances memory this provides evidence for the fact that info may be in there even if you cant retrieve it at the time of the test methodological: problems with concluding "no" memory - this doesn't mean it hasn't been encoded or stored - there could be something with retrieval of person is failure to perform about failure to encode or inability to retrieve theoretical: is everything still there?? But we just cant retrieve it... we don't have the right retrieval cues.

Overall Conclusions - real world memory: take interference and substantial delay for granted - having other events happens is taken for granted, there are some things we can do to improve memory how to improve memory? - best to semantically elaborate material at study - give meaning, integrate it with what you already know, space your study, match retrieval context to study, generate distinctive cues at test. Memory accuracy in relation to standard memory disorders MEMORY DISORDERS - general point - nobody's memory is perfect - we will look at damage to brain - head injury - car accident etc - not progressive disorders - but we look at sharp injury or events. Temporary Disorders: Symptoms of these types of disorders Theoretically: is there a storage deficit? Or did they learn it and cannot retrieve it? This is good to be able to test for temporary disorders Mild head injury: - temporary loss of past memories when person wakes up amnesia - problem with memory retrograde amnesia - problem with past memories - forgetting events that occurred prior to injury, could remember them before injury, but cant since injury anterograde amnesia - fail to learn or retrieve new information, trying to learn things for the first time after injury cant be done - this is in respect to when injury occurred - old / new info - before / after injury mild injuries (knock you out for a couple of hours) - person has forgotten events for some time before injury - cant remember injury or things that happened on the day of injury, or even info that happened in the couple of days prior. - the worse the injury - the more time into the past the memory problems last - unconscious for two weeks - may be quite permanent - unconscious for 2 hours - going to be less permanent - length of coma - tends to be a good indicator of the severity of problems that will be associated with memory islands of memory returning - remember certain things - but cant connect - isn't placed properly in time. If mild (an hour) - perhaps after a few days will remember everything except for like 30 seconds or so before injury. - e.g. can never remember accident itself 'e This is known as retreating retrograde amnesia this indicates that this is retrieval deficit, rather than a storage deficit.

Amnesia is not due to damage to storage, but damage to ability to retrieve for a short time Very rare disorder - transient global amnesia (TGA) Become an amnesic for like 24 hours/8 hours etc PERHAPS -Vascular problem in brain - mini stroke - temporary blockage of blood etc. This symptoms for TGA are the same as for PGA (permanent) but in TGA they resolve Symptoms: - severe retrograde amnesia - may forget last 10 years of life, can remember up to certain period, e.g. may forget that you are married but don't forget childhood - anterograde amnesia - inability to learn info tricky to do tests on TGA - as they are in this state for a short period of time. Three different types of test - paired associates - recall story - recall diagram - draw it paired associates - in learning new material 2 groups 11 patients during TGA same subjects a month after TGA had gone away (learning new set of words control subjects all matched to clinical patients nothing intrinsically wrong after TGA episode terrible problems learning material during TGA episode recalling a story experiment given a mark for each core idea of story remembered delayed condition - filled delay - control 4/8 TGA after - 4/8 TGA during - close to zero Immediate - during tga its almost as good as controls and after TGA Diagram copy / recall - test of learning visual material Problem with memory processes, not just lanaguge processing or visually processing. TGA terrible at delay Working memory component is fine Adding new info to LT memory and retrieving it is the problem??

After episode they go back to normal - severe AA during episode How do we test for RA?? Public events and personal events (standardized) Don't really have to compare to controls, we can compare them to themselves which provides some nice data. Week 6, Lecture 1 - 27/8 Properties of transient global amnesia Retrograde amnesia - forgetting of material that was known prior to head injury Memory for public events: who had been president of US Memory for personal events: marriage, graduation etc TGA - can compare subject while in amnesic state to same subjects performance after TGA has resolved. SEE HANDOUT Once TGA had resolved that could report 15 events (40% of these events they could recall during amnesic episode Graded retrograde amnesia - they are worse for the most recent things when in amnesic state (ie last two years) Summary: performance of TGA for retrograde amnesia - temporally grade retrograde amnsiea - older memories are more intact during amnesic episode, working memory is still intact.

We remember older memories better than more recent ones after: much better retrieval of that same information. Retrograde demonstrates deficit in retrieval, not damage in traces themselves It stops access to them, but not a deficit in storage - but also a deficit in new learning - memory for events within the TGA period never returns; suggest that it was never encoded / stored. Things learnt during TGA episode is never recalled. (although this is only tested with explicit memory tests) - defintley a retrieval deficit - may also be a storage / encoding deficit. Note: so far these TGA patients (Which are hard to get hold of) have only been tested with explicit memory tests - what about implicit?? Perhaps it can only be recalled unconsciously. This has never been examined, for normal amnesia this is the case.

TGA People assume its something to do with a mini stoke, blockage of blood flow, doesn't predict future stroke. It is a bit of a mystery. Permanent organic disorders (1) classic amnesic disorders aka - "global amnesia" - problems with all domains of memory - personal events, public events, pictures, memory for words etc general symptoms: almost always include anterograde amnesia (AA) - learning new stuff sometimes include AA and RA - looks exactly like symptoms for TGA but permanent - very rarely, RA occurs alone - cant remember past life, but can learn new stuff can hold conversation about something neutral but if the conversation involves retrieval from long term memory they struggle causes: stokes, head injury, encephalitis, anoxia (cardiac arrest, lack of oxygen), surgery, dementia (alzeihemers), Korsakov syndrome (alcohol abuse) - most common cause which brain regions are affected - medial temporal lobe (particularly hippocampus) and di encephalic structures (thalamus, hypothalamus) - in pure cases, all other non-memory functions intact - patient HM. Surgery for epilepsy in 1953 - has awareness of deficit. Describes life "like waking up from a dream... everyday is alone in itself, whatever joy I've had and whatever sorrow I've had" - it is strange to have awareness of problem - Jimmie G (Sacks, 1985) - Korsakov syndrome was in US navy until 1965, perfectly competent. Heavy drinker.

Developed occasional memory problems in the late 60's, but "able to cope". Had no insight into his problems. In late 1970, "blew his top" became delirious and excited, admitted to hospital. (last 2-3 weeks) patients confabulate a lot. - seen by Sacks in 1975: 49 yrs old - had memory problems for 4 years since onset of Korsakov syndrome. This delirious state resolves, but now become aware he has forgotten parts of previous life and cant retain new information. In Korsakov syndrome other areas of brains affected, but the sever amnesia is the most striking symptom of the syndrome - social skills intact - if it has dropped out of working memory its not there anymore Formal testing of amnesics - short term / working memory intact - long term memory deficit - for those with RA (most do), temporally graded (e.g. lost last 20 yrs, but childhood memory normal) often it is quite gradual progression - For AA, apparently cant learn anything new. In dense amnesics this deficit very severe - e.g. complete failure to learn animal-name association in 4 weeks of practice (Cath Haslam) - tested sever amnesics and tried to get them to learn novel associations - very rare animals - tried to train them to explicitly produce name when shown the picture, they showed pretty much zero ability to do this - HOWEVER this is only true if conscious recollection required i.e. explicit memory if try to retrieve ifo implicitly, you can show that they have perfectly normal memory - preservation of some aspects of LTM - they have implicit memory intact and skill learning intact skill learning (procedural).

HM learned pursuit rotor (manual tracking of moving target). Could learn a manual control task at the same level as normals, took him a little longer to learn, but learnt at the same level as normal. ALSO: Learned mirror drawing at same rate as normals Week 6, Lecture 2 - 28/8 Normal memory test - learn words of pictures, then do a variety of tests (explicit - recall, cued, old / new - amnesic performance here is poor) but then test this implicitly - repetition priming Implicit memory: intact repetition priming. Graf, Mander and Squire (1984) Studied word list.

Amnesic memory poor in three explicit tests, but normal in stem completion (CAL with first word that comes to mind) 'a this is an important instruction "first word that comes to mind" use stopwatch to really make them thinking of first word. amnesics - very poor at free recall - cued recall - also very poor (cued is easier than free recall - recognition memory test - old new - easier again - amnesics again are severely impaired at this. - critical new test - used a test called stem completion 'e test of implicit retrieval 'e can add new info to LTM that can be retrieved implicitly 'e memory performance for amnesics for implicit retrieval is perfectly normal take home message: can add new info to LTM for amnesics problems with amnesics are severe, but are limited in scope - only problems with LT explicit memory Theoretical implications 1. problems are only with LT explicit memory: implicit and working memory intact 2. severe AA suggests encoding / storage deficit BUT implicit tests show have been able to store new info 'e an explicit retrieval deficit 3. again, are all memory failures retrieval deficits? Is every event that has ever happened to us actually stored? And we just cant retrieve it when we need to, this is a big call, but it could infact but true. The Korsakov Jimmie - has a very brilliant recollection of being 19, better than a normal 49 year old, it became uncovered Practical: severe amnesics are very impaired in their day to day life, potentially if you could show this info was infact stored, it is a matter of finding out how to get it out...

Rehabilitation of memory disorders - no success in rehabilitating lost explicit memory as a general ability (ie unable to reestablish conscious recollection - how to help the helper - the carer - as often those with amnesia have no awareness of problems - need to work on residual memory abilities e.g. Skill based learning; or to improve use of diaries. You can learn explicitly via practice, but if you move amnesic to a new hospital and need to train them to get to the new kitchen - this is often not successful, they cant do it when trainer is not present, the learning they have achieved in very specific it doesn't generalize very well - often not successful - training generalizes poorly (find room ok with trainer present, but not alone) - forget to look at diary Why cant they do this? To improve their everyday life?? - ONE REASON - without explicit memory we have difficulty learning from our mistakes.

We can consciously remember when something we did last time led to the wrong conclusion. Tend to strengthen previous erroneous response - Errorless learning: Baddeley and Wilson (1994) - train patient to learn something, so that they never make a mistake - suggested implicit learning intact, leads to repeating same error - how do we active this 'e don't let an error ever occur 'e give complete answer at first and gradually fade out cues 'e also called: method of fading cues every morning I must look at my diary every morning I must look at my every morning I must look etc 'e this is hard work FALSE MEMORIES "Accuracy / distortion " focus on LTM if we do produce a memory, is it necessarily accurate, it is a distortion more often than a complete fabrication. Back to talking about conscious recollection. There is a heap of evidence that when you are confident etc, then this does not necessarily mean that this is exactly what happened (1) context / schema /gist effects - memory you produce depends on context it was placed or based on some sort of schema (plan or list of standard things you expect to happen in a given circumstance) - Bartlett (1932) reconstructive memory - war of the ghosts story - asked oxford undergraduates - native American tribe story about religious beliefs etc. read them them story, later told them to retell sorry back to them, people tended to report back information that was more inline with their beliefs or the beliefs they held about what native Americans believed (stereotypes) - Pomp i and Lachlan (1967) - repeating back a story that you had heard, found a lot of recognition errors, thematic errors in prose recognition "chief resident jones" 'e no mention of the word doctor or surgeon they often said surgeon was explcitily stated and although it was implied it wasn't actually said. Also "ugly growth too large for removal" this was interpreted as cancer by subjects. Don't remember exact copy of what occured - our lab remember meaning not form example of distortion from context Week 7 - Lecture 1 False Memory Accuracy / distortion focus on LTM If we do produce a memory is it necessarily accurate?

(2) source monitoring - actually heard word? Or imagined it being said? - Unconscious plagiarism - forgetting who told you the gossip Johnson, Foley (1988) Expt 1: ss heard some words said by experimenter and imagined themselves saying the words 'e good at discriminating whether they had heard word or imagined word 'e not very many source monitoring errors when you are imagining yourself saying the words expt 2: ss heard the same words (1/2) by experimenter and imagined the experimenter saying the other half of the words 'e very poor heard / imagined discrimination at test 'e source monitoring effects, confusion in source when you imagine experimenter saying words 'e internal vs. external More recent source monitoring experiments 'a Rodger - looked at discrimination b / w imagining self doing so meth.