COGNITIVE PSYCHOLOGY
COGNITIVE PSYCHOLOGY MEMORY
THE TRADITIONAL VIEW OF MEMORY : ‘MULTISTORE Mview ODEL’ (ATKINSON & SHIFFRIN, 1968) The traditional of memory: •
Theories consider both the architecture (way the system is & organized) and t1968) he processes (activities ‘Multistore model’ (Atkinson Shiffrin, occurring within the system). They often are based around encoding, storage, and retrieval
SENSORY STORES Eg Iconic memory
ATTENTION
DECAY
SHORT-TERM MEMORY
DISPLACEMENT
REHEARSAL
LONG-TERM MEMORY
INTERFERENCE
NB: “Box and arrow model”: Boxes= memory stores; arrows = processes
• •
Traditional view: short term processes very different from long term ! multi-‐store models (as opposed to unitary-‐store models where this distinction is less clear) Structural view of memory: Short-‐term and long-‐term memory rely on separate memory systems with different properties (structures that have different qualities, may exist in different parts of the brain)
DEFINING MEMORY: KINGS OF KNOWLEDGE •
ge
• •
st the
ther
ing…
)
7/30/2013
Sensory memory: a brief literal copy of an event – The traditional view of memory: o Iconic memory: Iconic memory involves the memory of visual stimuli. It is a type of sensory ‘Multistore model’ (Atkinson & Shiffrin, 1968) memory that lasts very briefly before quickly fading. Sperling’s findings suggest that iconic memory lasts only 0.5 seconds. Others suggest this may be an underestimate (more like 1600ms) ATTENTION REHEARSAL o Echoic memory: auditory stimuli. “SHORT-TERM Playback facility” typically 2-‐4 LONG-TERM seconds. SENSORY STORES EgmIconic memory Short term emory: “Buffer” for temporary MEMORY maintenance of information MEMORY Long-‐term memory o Facts/concepts: e.g. Who wrote Romeo & Juliet? Semantic ! Abstract conceptual knowledge is independent of how/when/whether memory was INTERFERENCE stored DECAY DISPLACEMENT o Episodes/events: e.g. Where did you go to school? Episodic ! Autobiographical, specific 7/30/2013 memory of particular time and place. You were an active participant. Structural view of memory: o Procedures: e.g. Driving a car, typing, reading, recognizing faces. Procedural ! knowledge of Short-term and long-term memory rely on separate memory how systems to do things, often unable to be verbalized or consciously accessed with different properties
TRADITIONAL VIEW OF TYPES OF KNOWLEDGE: STM VC LTM
Traditional view of types of knowledge: STM vs LTM SHORT TERM MEMORY (STM) Capacity
Limited 7+ 2
LONG-TERM MEMORY (LTM) Unlimited
Short-term memory (STM): Decays within 30 Forgetting due to seconds if not interference Letter span tasks rehearsed rather than decay
Rate of forgetting
• How many items do people normally Type of code Phonological Semantic remember? • What if they have to count out loud while remembering?
SHORT-‐TERM MEMORY (STM): LETTER SPAN TASKS • • • • •
How many items do people normally remember? 7 +/-‐ 2 What if they have to count out loud while remembering? Reduces ability to rehearse the letters What if letters sound the same? The sounds get confused within the rehearsal process What if words instead of letters? Still 7 +/-‐ 2. Amount of syllables would make a difference though. Simon (1974) found that the span in chunks was less with larger chunks than smaller
MULTI-‐STORE MODEL: ENCODING PROCESS • •
•
How does information get into LTM? Traditional view: o It must be attended to and registered in iconic memory o REHERSAL required to maintain/transfer to LTM o Repetition/rehearsal critical to effective memory (Ebbinghaus, 1885 – list of nonsense words) o But how/why does this facilitate encoding? " Repetition alone not sufficient (Craik & Watkins, 1973 – people presented with words one at a time, their task to remember the last word that started with G) " Memory better when people rehearse silently rather than out loud (Kellas et al. 1975 – compared memory performance between groups – memory was better in the group that recited silently) " Not JUST rehearsal Levels of processing hypothesis (Craik & Lockhart, 1975) – critical phenomena in memory o Investigated incidental memory (rather than intentional memory) for material presented with different “orienting tasks” (weren’t told that they were going to be tested on memory) " How many letters does this word have? – (Structural) shallow processing " Does this word rhyme with treat? (Phonological task) " Do you find these in a city? (Semantic task) – deep processing o Memory best when encoding task requires deep rather than shallow processing o NOT due simply to longer encoding time for deeper encoding tasks o ENCODING DOES NOT DEPEND SOLELY ON REHEARSAL o BUT how do we know if orienting task requires “deep processing”? o See whether it yields better memory o CIRCULAR?!
WHAT DEFINES “DEEP” PROCESSING? •
•
•
Memory is better for material that is semantically organized at encoding o By requiring categorisation at encoding o When more elaboration is encouraged during encoding o When material is self-‐relevant But why does deep processing help? o Depth of processing describes rather than explains effectiveness of memory o What is the form of existing knowledge? o How is knowledge represented in the mind (brain)? o How is this knowledge retrieved? BUT deep processing does not ALWAYS yield better memory
RETRIEVAL PROCESSES •
•
• •
How do we retrieve information encoded in LTM? o Explicit voluntary retrieval e.g. memory search o Memory activated by retrieval cue: familiar; “just know” -‐ e.g. friends name Measuring memory o Free recall task: report items from earlier study episode o Recognition task: Select previously studied items from mixture of old and new items o Recognition (nearly always) better than recall Why? Recall task does not provide any (explicit) cue Recognition task provides a cue (the studied item) which can activate (“prime”) memory network
‘Transfer appropriate processing’ principle (Morris, Bransford & Franks, 1977
Memory network can also be primed by the study context e.g. Godden & Baddeley (1975)* Encoding specificity principle (Tulving & Thomson, 1973) Transfer appropriate processing: Memory performance usually best for “deeply” encoded items. BUT Retrieval is best when there is a match between the also depends on the “match” between the processes required at encoding and retrieval (Morris et processes al., 1977)* required at encoding and retrieval Adapted from Morris, Bransford & Franks (1977) “shallow” processing yields 0.9 better memory 0.8 WHEN it matches 0.7 task context 0.6 Proportion recognized
• • •
0.5 0.4 0.3 0.2
0.1
Standard test Rhyming test
Standard: did you see it before? Rhyme Semantic Rhyming: did you Orienting task see a word that EAGLE: Does it rhyme with legal? Is it a bird rhymes with regal? o When semantic encoding, best cues relate to meaning, not surface form. E.g. ‘The fish attacked the swimmer’ ! shark a better retrieval cue than fish Godden & Baddeley (1975): Participants learned words either on land or underwater; and were asked to later recall the words on land or underwater o Recall best when the contexts matched o Recognition unaffected by context o Different retrieval tasks probe different aspects of memory o Evidence for the encoding specificity principle – what we encode is specific to the context in which we learn it 0
•
EXPLICIT VS IMPLICIT MEMORY TASKS •
•
Traditional approach: Explicit memory tasks o e.g. Recall/Recognition o Subjects explicitly told to remember items from previous list ! intentional retrieval o Memory performance shows levels of processing and transfer-‐appropriate processing effects Implicit memory tasks o NOT told to try to remember, just to perform a task ! incidental o e.g. fragment completion, stem completion, perceptual identification ! compare performance for old and new items to infer memory o Repetition priming also an index of implicit memory
STRENGTHS OF MULTI-‐STORE MODEL •
•
Traditional information processing approach o Highlights conceptual distinctions between different types of memory that vary in temporal duration, storage capacity, forgetting mechanisms o Consistent with evidence of double dissociations between the effects of brain damage on STM and LTM (amnesia – LTM affected but not STM) Central assumptions o Cognitive processing consists of transforming information into “higher”, more durable forms of memory o ! Processing is “bottom up”: low level (perceptual) processing precedes high level (cognitive) processing o ! attention necessary to “select” information from sensory memory and for high level processing
7
THE TRADITIONAL VIEW OF MEMORY : ‘MULTISTORE Mview ODEL’ (ATKINSON & SHIFFRIN, 1968) The traditional of memory: •
Theories consider both the architecture (way the system is & organized) and t1968) he processes (activities ‘Multistore model’ (Atkinson Shiffrin, occurring within the system). They often are based around encoding, storage, and retrieval
SENSORY STORES Eg Iconic memory
ATTENTION
DECAY
SHORT-TERM MEMORY
DISPLACEMENT
REHEARSAL
LONG-TERM MEMORY
INTERFERENCE
NB: “Box and arrow model”: Boxes= memory stores; arrows = processes
• •
Traditional view: short term processes very different from long term ! multi-‐store models (as opposed to unitary-‐store models where this distinction is less clear) Structural view of memory: Short-‐term and long-‐term memory rely on separate memory systems with different properties (structures that have different qualities, may exist in different parts of the brain)
DEFINING MEMORY: KINGS OF KNOWLEDGE •
ge
• •
st the
ther
ing…
)
7/30/2013
Sensory memory: a brief literal copy of an event – The traditional view of memory: o Iconic memory: Iconic memory involves the memory of visual stimuli. It is a type of sensory ‘Multistore model’ (Atkinson & Shiffrin, 1968) memory that lasts very briefly before quickly fading. Sperling’s findings suggest that iconic memory lasts only 0.5 seconds. Others suggest this may be an underestimate (more like 1600ms) ATTENTION REHEARSAL o Echoic memory: auditory stimuli. “SHORT-TERM Playback facility” typically 2-‐4 LONG-TERM seconds. SENSORY STORES EgmIconic memory Short term emory: “Buffer” for temporary MEMORY maintenance of information MEMORY Long-‐term memory o Facts/concepts: e.g. Who wrote Romeo & Juliet? Semantic ! Abstract conceptual knowledge is independent of how/when/whether memory was INTERFERENCE stored DECAY DISPLACEMENT o Episodes/events: e.g. Where did you go to school? Episodic ! Autobiographical, specific 7/30/2013 memory of particular time and place. You were an active participant. Structural view of memory: o Procedures: e.g. Driving a car, typing, reading, recognizing faces. Procedural ! knowledge of Short-term and long-term memory rely on separate memory how systems to do things, often unable to be verbalized or consciously accessed with different properties
TRADITIONAL VIEW OF TYPES OF KNOWLEDGE: STM VC LTM
Traditional view of types of knowledge: STM vs LTM SHORT TERM MEMORY (STM) Capacity
Limited 7+ 2
LONG-TERM MEMORY (LTM) Unlimited
Short-term memory (STM): Decays within 30 Forgetting due to seconds if not interference Letter span tasks rehearsed rather than decay
Rate of forgetting
• How many items do people normally Type of code Phonological Semantic remember? • What if they have to count out loud while remembering?
SHORT-‐TERM MEMORY (STM): LETTER SPAN TASKS • • • • •
How many items do people normally remember? 7 +/-‐ 2 What if they have to count out loud while remembering? Reduces ability to rehearse the letters What if letters sound the same? The sounds get confused within the rehearsal process What if words instead of letters? Still 7 +/-‐ 2. Amount of syllables would make a difference though. Simon (1974) found that the span in chunks was less with larger chunks than smaller
MULTI-‐STORE MODEL: ENCODING PROCESS • •
•
How does information get into LTM? Traditional view: o It must be attended to and registered in iconic memory o REHERSAL required to maintain/transfer to LTM o Repetition/rehearsal critical to effective memory (Ebbinghaus, 1885 – list of nonsense words) o But how/why does this facilitate encoding? " Repetition alone not sufficient (Craik & Watkins, 1973 – people presented with words one at a time, their task to remember the last word that started with G) " Memory better when people rehearse silently rather than out loud (Kellas et al. 1975 – compared memory performance between groups – memory was better in the group that recited silently) " Not JUST rehearsal Levels of processing hypothesis (Craik & Lockhart, 1975) – critical phenomena in memory o Investigated incidental memory (rather than intentional memory) for material presented with different “orienting tasks” (weren’t told that they were going to be tested on memory) " How many letters does this word have? – (Structural) shallow processing " Does this word rhyme with treat? (Phonological task) " Do you find these in a city? (Semantic task) – deep processing o Memory best when encoding task requires deep rather than shallow processing o NOT due simply to longer encoding time for deeper encoding tasks o ENCODING DOES NOT DEPEND SOLELY ON REHEARSAL o BUT how do we know if orienting task requires “deep processing”? o See whether it yields better memory o CIRCULAR?!
WHAT DEFINES “DEEP” PROCESSING? •
•
•
Memory is better for material that is semantically organized at encoding o By requiring categorisation at encoding o When more elaboration is encouraged during encoding o When material is self-‐relevant But why does deep processing help? o Depth of processing describes rather than explains effectiveness of memory o What is the form of existing knowledge? o How is knowledge represented in the mind (brain)? o How is this knowledge retrieved? BUT deep processing does not ALWAYS yield better memory
RETRIEVAL PROCESSES •
•
• •
How do we retrieve information encoded in LTM? o Explicit voluntary retrieval e.g. memory search o Memory activated by retrieval cue: familiar; “just know” -‐ e.g. friends name Measuring memory o Free recall task: report items from earlier study episode o Recognition task: Select previously studied items from mixture of old and new items o Recognition (nearly always) better than recall Why? Recall task does not provide any (explicit) cue Recognition task provides a cue (the studied item) which can activate (“prime”) memory network
‘Transfer appropriate processing’ principle (Morris, Bransford & Franks, 1977
Memory network can also be primed by the study context e.g. Godden & Baddeley (1975)* Encoding specificity principle (Tulving & Thomson, 1973) Transfer appropriate processing: Memory performance usually best for “deeply” encoded items. BUT Retrieval is best when there is a match between the also depends on the “match” between the processes required at encoding and retrieval (Morris et processes al., 1977)* required at encoding and retrieval Adapted from Morris, Bransford & Franks (1977) “shallow” processing yields 0.9 better memory 0.8 WHEN it matches 0.7 task context 0.6 Proportion recognized
• • •
0.5 0.4 0.3 0.2
0.1
Standard test Rhyming test
Standard: did you see it before? Rhyme Semantic Rhyming: did you Orienting task see a word that EAGLE: Does it rhyme with legal? Is it a bird rhymes with regal? o When semantic encoding, best cues relate to meaning, not surface form. E.g. ‘The fish attacked the swimmer’ ! shark a better retrieval cue than fish Godden & Baddeley (1975): Participants learned words either on land or underwater; and were asked to later recall the words on land or underwater o Recall best when the contexts matched o Recognition unaffected by context o Different retrieval tasks probe different aspects of memory o Evidence for the encoding specificity principle – what we encode is specific to the context in which we learn it 0
•
EXPLICIT VS IMPLICIT MEMORY TASKS •
•
Traditional approach: Explicit memory tasks o e.g. Recall/Recognition o Subjects explicitly told to remember items from previous list ! intentional retrieval o Memory performance shows levels of processing and transfer-‐appropriate processing effects Implicit memory tasks o NOT told to try to remember, just to perform a task ! incidental o e.g. fragment completion, stem completion, perceptual identification ! compare performance for old and new items to infer memory o Repetition priming also an index of implicit memory
STRENGTHS OF MULTI-‐STORE MODEL •
•
Traditional information processing approach o Highlights conceptual distinctions between different types of memory that vary in temporal duration, storage capacity, forgetting mechanisms o Consistent with evidence of double dissociations between the effects of brain damage on STM and LTM (amnesia – LTM affected but not STM) Central assumptions o Cognitive processing consists of transforming information into “higher”, more durable forms of memory o ! Processing is “bottom up”: low level (perceptual) processing precedes high level (cognitive) processing o ! attention necessary to “select” information from sensory memory and for high level processing
7
