Joel Cooper

1
Memory
Joel Cooper University of Utah
2
Memory Is

• The mechanism we use to create, maintain and
  retrieve information about the past

3
Processes in Memory

• Encoding
• Processes used to store information in memory
• Storage
• Processes used to maintain information in memory
• Retrieval
• Processes used to get information back out of
  memory

4
Methods Used to Study Memory

• Which type of memory test would you rather have?
• An essay or a multiple choice exam?
• recall vs. recognition

5
Recall Tasks

• Free Recall
• Recall all the words you can from the list you
  saw previously
• Cued Recall
• Recall everything you can that is associated with
  _______
• Participants are given a cue to facilitate recall
• Serial Recall
• Recall the names of all previous presidents in
  the order they were elected
• Need to recall order as well as item names

6
Recognition Tasks

• Circle all the words you previously studied
• Indicate which pictures you saw yesterday

7
Implicit or Explicit Memory Tasks

• Explicit memory tasks
• Involves conscious recollection
• Participant knows they are trying to retrieve
  information from their memory
• Implicit memory tasks
• Require participants to complete a task
• The completion of the task indirectly indicates
  memory

8
Implicit Memory Tasks

• Participants are exposed to a word list
• Tiger
• Lion
• Zebra
• Panda
• Leopard
• Elephant
• After a delay

• Participants then complete word puzzles, they are
  not aware they are a type of memory test
• Word fragment Completion
• C_E_TA_
• E_E_ _AN_
• _ E _ RA
• Word Stem Completion
• Mon _____
• Pan_____

9
Models of Memory

• Represent ways that memory has been
  conceptualized
• Atkinson Shiffrins 3 Stage Model of Memory
• Craik Lockharts Level of Processing Model
• Baddeleys Working Memory Model
• Tulvings Multiple Memory Systems Model
• McClelland Rumelharts Connectionist Model

10
Traditional Model of Memory

• Atkinson Shiffrin (1968) 3 Stage Model

Short Term Memory (STM)
Long Term Memory (LTM)
Sensory registers
Stimuli
Information Processing Model
11
Sperling (1960) Iconic Memory Research

• Whole report procedure
• Flash a matrix of letters for 50 milliseconds
• Identify as many letters as possible
• Participants typically remembered 4 letters
• Partial Report Procedure
• Flash a matrix of letters for 50 milliseconds
• Participants are told to report bottom row
• Participants were able to report any row
  requested

12
Whole Report

13
I G X E N R B P J Q K
C

14
(No Transcript)
15
I G X E N R B P J Q K
C

16
Partial Report Technique
High

Middle
Low
17


18
K O D G N E L S W A Q Y

19
(No Transcript)
20
K O D G N E L S W A Q Y

21
High
Middle

• Low

22

23
V L U G Q S F I M B O D

24
(No Transcript)
25
V L U G Q S F I M B O D

26
Partial Report Technique

27
Averbach Coriell (1961) Iconic Memory Research
G E U L M F S X W P M B D H J Y

• - Showed matrix for 50 msec
• - Place a small mark above a letter at different
  delays
• Results indicated that as many as 12 letters
  could be stored in
• sensory memory

28
Sensory Stores

• Iconic store or Visual sensory register
• Holds visual information for 250 msec longer
• Information held is pre-categorical
• Capacity up to 12 items
• Information fades quickly
• Econ or Auditory sensory register
• Holds auditory information for a 2-3 seconds
  longer to enable processing

29
Short-Term Memory
Rehearsal

• Attention
• Attend to information in the sensory store, it
  moves to STM
• Rehearsal
• Repeat the information to keep maintained in STM
• Retrieval
• Access memory in LTM and place in STM

Short Term Memory (STM)
Attention
Storage Retrieval
30
Short-Term Memory

• Short-term /working memory
• Limited capacity (7 or - 2) or 2 sec.
• Inputs from SR and LTS
• Consciousness
• Coding verbal/spatial
• Information can be maintained indefinitely,provid
  ed it is given constant attention
• Information decays in 15-20 seconds
• Rote vs Chunking mnemonics

Webster says Mnemonic Assisting or designed to
assist memory
31
Short-Term Memory

• How is information lost?
• Decay gt Time
• Interference gt Older displaced by new
• Brown-Peterson Task
• Waugh Norman

32
Brown-Peterson Task

• Subjects presented with trigram (XQJ)
• Experimenter presents number (257)
• Subject counts backwards by 3s (2/sec)
• After x seconds, subjects recall trigram

33
Brown-Peterson Task

• VRO
• 187

UYV 89
IDC 131
34
Brown-Peterson Task
35
Proactive Interference in STM
0

• Keppel Underwood (1962)
• Replicated the Peterson Peterson Task varying
  the time delay to recall
• Analysis was done by trial number (1st trial, 2nd
  trial, 3rd trial, etc.)
• Found support for proactive interference

36
Loss or Decay?

• Waugh Norman -- Loss in B-P task could be due
  to decay or interference, because both covaried
  with retention interval
• Independently varied time and number of
  intervening items
• Does loss follow time or number of items?

37
Waugh Norman Task

• Subjects presented with string of digits
• Digits read at 1 or 4 per second (time)
• Does recall vary as a function of time or items?

1 per second 6 2 9 4 1 8 3 4 8
1 0 4 2 6 7 3 ( 16 sec ) 4 per
second 6294183481042673 ( 4 sec )
Probe Digit 9 Response 4
38
Waugh Norman Task
39
Loss from STM

• Loss is largely due to interference
• Old information is replaced by new
• Rehearsal moves info to head of buffer

40
Long-Term Memory

• Capacity
• Thus far limitless
• Duration
• Potentially permanent

Long Term Memory (LTM)
41
Long-Term Store

• Information comes from STS
• Conscious transfer -- Explicit memory
• Unconscious transfer -- Implicit memory
• Large capacity
• Can information be lost from LTS?
• Poor retrieval cues
• Memories overwritten?

42
Lets Test Your LTM!

• You will see several words, one at a time
• Do whatever you can to try and remember as many
  of the words as you can
• At the end of the list, try to recall as many
  words as you can

43
Bed Clock Dream Night Turn Mattress Snooze Nod Nig
ht
Artichoke Insomnia Rest Toss Night Alarm Nap Snore
Pillow
44
Write down the words you saw
45
Words

• Bed
• Clock
• Dream
• Night
• Turn
• Mattress
• Snooze
• Nod
• Night

• Artichoke
• Insomnia
• Rest
• Toss
• Night
• Alarm
• Nap
• Snore
• Pillow

46
Memory Demonstration

• Did you Recall?
• Bed or Clock
• Snore or Pillow
• Night
• Artichoke
• Toss and Turn
• Sleep

• Explanation
• Primacy
• Recency
• Spacing Effect
• Distinctiveness
• Clustering
• False Memory

47
Serial Position Curve
48
Effects of Rehearsal

• Rhundus -- subjects rehearsed outloud
• Primacy curve matched rehearsal curve
• Primacy determined by transfer to LTM

49
Effects of Distractors

• Glanzer -- distractor task at end of list
• Recency portion of curve abolished
• Recency determined by readout from STM

50
Serial Position Curves

• Primacy determined by transfer to LTM
• Recency determined by readout from STM

51
Levels of Processing Model of Memory

• Craik Lockhart (1972)
• Process level different strengths of
  memories
• Deep processing better memory
  elaborating according to meaning leads to a
  strong memory
• Shallow processing emphasizes the physical
  features of the stimulus the memory trace is
  fragile and quickly decays
• Maintenance rehearsal (Rote) vs. Elaborative
  rehearsal

52
Support for Levels of Processing

• Craik Watkins (1973)
• Participants listened to lists of words
• Task was to recall the last word in the list
  which began with a particular letter
• The number of intervening words between words
  beginning with the target letter was varied

53
Craik Watkins (1973) Results

• Recall of words was independent of the length of
  time (the number of intervening words) it was
  maintained in STM
• Conclusion Maintenance rehearsal did not
  automatically lead to LTM
• Levels-of-Processing Interpretation Students
  rehearsed the words without elaborating on the
  meaning of the words, only concentrating on the
  initial consonant soundrehearsing at a shallow
  level

54
Depth of Processing

• Craik Lockhart LEVEL QUESTION-
  Structural Capital letters - Phonemic Rhyme
  - Categorical Type of fish - Sentence Fit
  in sentence

55
Craik Tulving (1975) Results
56
Criticisms of LOP Model

• Circular definition of levels
• Transfer appropriate processing effect
• Morris, Bransford, and Franks (1977)
• Two processing tasks semantic vs. rhyme
• Two types of tests standard yes/no recognition
  vs. rhyme test 
• Memory performance also depends on the match
  between encoding processes and type of test

57
Mnemonic Strategies

• Chunking
• Method of Loci
• Peg Word Mnemonic

58
Chunking

• Grouping elements into chunks
• Chase and Ericcsons subjects
• Chunked digits into running times
• Remembered up to 82 digits!
• Not as effective as more elaborative strategies

59
Method of Loci

• Based upon visual imagery
• Imagine a grocery list
• (eggs, milk, cheese, bread, butter)
• Imaging items placed in a common scene
• To recall, mentally stroll through scene
• Bizarreness / distinctiveness

60
Peg Word Mnemonic

• Uses prememorized list (e.g., rhyme)
• One is a bun gt
• Two is a shoe gt
• Three is a tree gt
• Four is a door gt

61
Acronyms

• Create word from beginning letters
• Create a limerick from beginning letters
• OOOTTAFAGVAH 12 Cranial nerves

62
Why Do Mnemonics Devices Work?

• Provide structure for learning
• Provide durable trace (less interference)
• Provide retrieval structure

63
Which Mnemonic is the Best?

• Roediger (1980)

64
Working Memory Model Baddeley Hitch, 1974
Phonological Loop
Visuospatial Sketchpad
Central Executive
Spatial - Where
Visual - What
Phonological store - 2 sec.

• Norman and Shallice 1986
• Action Plan
• Contention Scheduling
• Supervisory attentional system (SAS)

Articulatory control process
65
Working Memory Model

• Articulatory Loop
• Used to maintain information for a short time and
  for acoustic rehearsal
• Visuo-spatial Sketch Pad
• Used for maintaining and processing visuo-spatial
  information
• Episodic Buffer
• Used for storage of a multimodal code, holding an
  integrated episode between systems using
  different codes

66
Working Memory Model

• Central Executive
• Focuses attention on relevant items and
  inhibiting irrelevant ones
• Plans sequence of tasks to accomplish goals,
  schedules processes in complex tasks, often
  switches attention between different parts
• Updates and checks content to determine next step
  in sequence of parts

67
Working Memory Model Support

• Baddeley (1986)
• Participants studied two different list types
• 1 syllable wit, sum, harm, bay, top
• 5 syllables university, opportunity, aluminum,
  constitutional, auditorium
• Reading rate seemed to determine recall
  performance
• Supports conceptualization of an articulatory loop

68
Working Memory Model Support

• Visuo-spatial Sketch Pad
• Dual-task paradigm
• Sketchpad can be disrupted by requiring
  participants to tap repeatedly a specified
  pattern of keys or locations while using imagery
  at the same time

69
Multiple-Memory Systems Model

• Tulving (1972)
• Semantic Memory
• General knowledge
• Facts, definitions, historical dates
• Episodic Memory
• Event memories (first kiss, 6th birthday)
• Procedural Memory
• Memories on how to do something (skiing, biking,
  tying your shoe)

70
Multiple-Memory Systems Model Support

• Nyberg, Cabeza, Tulving (1996)
• PET technology to look at episodic and semantic
  memory
• Asked people to engage in semantic or episodic
  memory tasks while being monitored by PET  

• Results 
• Left (hemisphere) frontal lobe differentially
  active in encoding (both) and in semantic memory
  retrieval
• Right (hemisphere) frontal lobe differentially
  active in retrieval of episodic memory

71
Connectionist Perspective

• Parallel distributed processing model
• Memory uses a network
• Meaning comes from patterns of activation across
  the entire network
• Spreading Activation Network Model
• Supported by priming effects

72
Deficient Memory

• Amnesias
• Retrograde Amnesia
• Loss of memory for events that occurred before
  the trauma
• Infantile Amnesia
• Inability to recall events of young childhood
• Antereograde Amnesia
• No memory for events that occur after the trauma

73
Amnesia Studies

• Study antereograde amnesiacs using implicit and
  explicit memory tests
• Amnesiacs show normal priming (implicit), but
  poor recognition memory (explicit)
• They did not remember having seen the word list,
  but completed the word fragments at the same rate
  as normals

74
Hippocampus and Memory

• Hippocampus
• Critical for integration and consolidation
• Essential for declarative memory
• Without the hippocampus only the learning of
  skills and habits, simple conditioning, and the
  phenomenon of priming can occur

75
Interaction of Encoding and Retrieval Processes

• Anderson Pichert (1978)
• Participants read a story about the activities of
  two boys at home either from the point of view of
  a burglar or a homebuyer
• Later the participants were asked to remember as
  much as possible about the story they read
• The point of view affected what participants
  recalled
• Homebuyers were more likely to remember that the
  basement was musty
• Burglars were more likely to remember the coin
  collection and color TV

76
To Understand Memory Processes

• Need to understand encoding processes, retrieval
  processes, and how they function together for
  short term storage (STM) and long term storage
  (LTM)

77
Encoding Processes

• Creating an acoustic code
• What it sounds like
• Creating a semantic code
• What it means
• Creating a visual code
• What it looks like

78
Encoding Types and STM

• Type of code may rely on type of task
• STM refers to memory that needs to be held
  temporarily
• Evidence exists for a variety of encoding types
  for STM

79
Evidence for Acoustic Encoding in STM

• Conrad (1964)
• Visually present a series of letters
• Ask participants to write the order letters are
  presented
• What types of errors are made?

80
Conrad (1964)

• Found evidence for the use of an acoustic code in
  STM
• Participants made acoustic errors
• F for S, B for V, P for B
• Not visual errors
• E for F, O for Q, R for P
• Participants encoded items acoustically even
  though stimuli were presented visually

V . . . B!
81
Shulman (1970)

• Evidence for semantic encoding in STM
• Participants viewed 10-word lists
• Given a recognition test using visually
  represented "probe words" which were either
• Homonyms - e.g. "bawl" for "ball"
• Synonyms - e.g. "talk" for "speak"
• Identical to the original word

82
Shulman (1970) Results

• The Homonym and Synonym probes produced similar
  error rates - this suggests that an equal amount
  of acoustic and semantic processing must be
  taking place
• Homonyms - e.g. "bawl" for "ball"
• Synonyms - e.g. "talk" for "speak"
• Identical to the original word

83
Posner Keele (1967)

• Evidence for visual encoding in STM
• Letter matching task
• Two letters separated by brief interval
• Participant had to indicate if same letter
• A-a Yes
• A-A Yes
• A-M No
• Measure reaction time

84
Posner Keele (1967) Results

• If letters were the same visually (a-a)
  participants were faster than if the letters were
  not the same visually (A-a)
• Results indicate that visual code was also
  present for STM

85
Encoding Types LTM

• Type of code may rely on type of task
• LTM refers to memory that may be held permanently
  
• Evidence exists for a variety of encoding types
  for LTM

86
Semantic Encoding in LTM

• Grossman Eagle (1970)
• Study 41 different words
• Given recognition test after delay
• 9 of the distractors were semantically related to
  words on list
• 9 of the distractors were not
• False alarms for each type 1.83 of synonyms, but
  only 1.05 of unrelated

87
Visual Encoding in LTM

• Frost (1972)
• Participants studied 16 drawings
• Manipulated visual orientation and semantic
  category
• After a delay, participants were asked if they
  had studied an object with the same name as the
  test object
• Reaction time was measured
• Participants responded faster to identical
  drawings than drawings in a different orientation
• This result indicates visual encoding occurred

88
Acoustic Encoding in LTM

• Evidence of very long-term memory for songs
• Rubin (1977)
• Participants recall more of the text when
  provided with the melody of a well-learned song
  ("Star Spangled Banner") than when given no cue

89
Transfer from STM to LTM

• Consolidation
• Integrating new information into stored
  information
• Disruption of consolidation is studied in
  amnesiacs
• ECT patients (Squire)

90
Principles to Strengthen Memory

• Elaborative rehearsal is better than maintenance
  rehearsal
• Distributed practice is better than massed
  practice
• Spacing effect
• Organizing information to enhance memory

91
Why Does Distributed Practice Work?

• REM Theory
• The more REM sessions following study sessions,
  the more consolidation that occurs
• Multiple encoding contexts theory
• Multiple study sessions lead to multiple types of
  encoding, thus more possibility of matching
  during test conditions

92
Prospective Memory

• The ability to remember a future intention
• Buying bread on your way home from work
• Going to the dentist on Wednesday
• Retrospective memory is memory of the past

93
Retrieval Processes

• Getting information back out
• Multiple processes can be used to enhance
  retrieval
• Different strategies are used for short term
  storage and long term storage
• Matching the type of processes done during
  encoding with the type of processes done at
  retrieval increases success

94
Retrieval from STM

• Is the search serial or parallel?
• Serial indicates one by one search
• Parallel means all items are processed at once
• Is the search exhaustive or self-terminating?
• Exhaustive indicates that all items in the set
  are examined
• Self-terminating means that after target is found
  the search stops

95
Studying Searching in STM
0

• Saul Sternberg (1967)
• Memorize a set of numbers (6,3,8,2,7)
• Shown a probe digit
• Participant must indicate if the probe was in the
  set
• Reaction time to respond is measured

2
6,3,8,2,7
Yes
96
Sternberg (1967)
0

• 3 critical factors manipulated
• How many items were in the set the participants
  had to memorize
• Whether the probe was in the list
• The probes location in the set

97
Sternberg (1967)

• Possible Result Patterns
• A represents parallel processing
• B illustrates serial processing
• C illustrates exhaustive serial processing
• D illustrates self-terminating serial processing

98
Sternbergs Conclusion

• A serial exhaustive model
• But.
• Corcoran (1971) proposed that a parallel model
  could also explain the pattern found
• Townsend (1971) stated it was mathematically
  impossible to distinguish parallel from serial
• Thus, both models still exist

99
If You Do Not Retrieve from LTM

• Has the memory disappeared?
• or
• Is the memory still there but cannot retrieve it
  (available, but not accessible)?

100
Evidence Supporting Still There Theory Nelson
(1971)

• Paired associate List
• 43-house
• 67-dog
• 38-dress
• 77-sissors
• Cued recall test
• 43- ________
• 67- ________

• Two week delay
• Subjects recalled 75 of items on list
• But focus was on 25 they forgot.

101
Nelson (1971) Critical Manipulation

• If participants forgot 38-dress and
  77-sissors then participants relearned either
  same pairs or changed pairs

The better performance of participants in the
same condition indicate that there was some
memory left for forgotten items. Otherwise
both groups would remember the same amount.
102
What Contributes to Forgetting?

• Decay theory
• Memory is weakened with disuse
• Interference theory
• Proactive old memories interfere with recall of
  new information
• Retroactive new memories interfere with recall
  of old information

103
Retroactive Interference from LTM
0
The experimental group will remembers less
material from the tested list A compared to the
control group Information learned afterwards
interferes with retrieval of List A.
104
Proactive Interference from LTM
The experimental group remembers less material
from the tested list B than the control
group Information previously learned (list A)
interferes with retrieval of List B
105
Flashbulb Memories

• Some researchers propose that events that are
  particularly surprising or arousing will yield
  flashbulb memories
• Where were you when the
• Challenger explosion occurred?
• OJ verdict was read?
• JFK was assassinated?
• Bombing of the twin towers?

106
(No Transcript)
107
Flashbulb Memories

• Some research proposes good memory for
• Place where you learned of information
• What you were doing when you heard it
• Where you heard the information from
• Emotions in self and others
• The aftermath

108
Emotion and Memory

• There is a strong relationship (.90) between the
  emotionality and vividness of memory
• This does not mean that the memory is accurate
• Emotional events seem to be less resistant to
  forgetting over time
• Perhaps they are perceived better
• Perhaps we think about them more

109
Flashbulb Memory Results

• Neisser and Harsch (1992)
• Tested immediate memory for Shuttle Explosion,
  and then tested it again 3 years later
• There was little agreement with the two
  memories despite the confidence of the
  participants

110
Encoding Specificity

• Memory is improved when information available at
  encoding is also available at retrieval

111
Encoding Specificity

• Tulving (1983)
• People encode the context with the target
  material
• Physical match (class, diving, smell)
• Emotional match (happy, depressed)
• Understanding match (childhood amnesia, under the
  influence of drugs match)

112
State Dependent Learning
113
State Dependent Learning
114
State Dependent Learning
115
State Dependent learning, why is it important?

• Content addressable memory
• Partial information helps in retrieval
• Adaptive function of memory
• Most relevant memories are most accessible