Transcript Chapter 7

Cognitive Processes
PSY 334
Chapter 7 – Human Memory:
Retention and Retrieval
What is Forgetting?
 Do memories still exist in mind when we
cannot remember?


Penfield – stimulated areas of the brain and
got reports of recall from childhood.
No way to check the accuracy of reports.
 Nelson – some savings are evident even
when subjects cannot remember items:

Savings found with recall (78% unchanged,
43% changed) and recognition (34% & 19%).
Brain Areas Important to
Memory
The Retention Function
 Wickelgren – studied the retention
function:



Performance is a function of delay.
Log (d’) = A – b log T
Where: T is delay, d’ is performance
(memory strength).
 Power law of forgetting -- power function
becomes linear when plotted on log-log
scales.
Power Law of Forgetting
Rate of Forgetting
 Retention function shows diminishing
loss (forgetting) with delay.
 Theory of short-term memory predicts
sharp drop-off followed by stable
memory.


Since all retention functions are like this,
there is nothing special about short-term
memory compared to long-term memory.
Practice postpones the point of decay.
Long-Term Retention
 Bahrick – studied retention of English-
Spanish vocabulary over 50 years.



Substantial practice effect.
Slow decline after 3 yrs.
Drop-off at end due to physical aging.
 Barnes – decrease in long-term
potentiation with delay.


Mirrors retention function.
Decay theory of forgetting – LTP changes.
Bahrick’s Retention of Spanish
Notice that
retention only
drops off at the
end when
subjects are 65+
in age.
Decay – Decline in LTP
Interference
 Interference paradigm – two groups
defined:


Experimental group – learns new
associations for previously learned list
Control group – learns entirely new list
 Typically the experimental group does
worse after a delay.
 Does this mean that it is difficult to
maintain multiple associations?
Typical Interference
Experiment
A-D Experimental
C-D Control
Learn A-B
Learn A-B
Learn A-D
Learn C-D
Test A-B
Test A-B
A-D produces more
interference than C-D
Fan Effect
 There is a limit to how much activation
can spread within a network:

The more associations, the less activation
can spread to any particular structure.
 Anderson – fan effect:


Recognition time increases with the
number of facts about a person and a
location (e.g., lawyer, church).
The brain works harder with “high” fans.
Anderson’s Experiment
Mean
Recognition
Time for Different
Numbers of
Sentences
1
2
1
1.11
1.17
2
1.17
1.22
Preexisting Memories
 Does knowledge brought into an
experiment interfere with new learning?
 Lewis & Anderson – facts about
Napoleon:



Fantasy facts – learned during experiment
True facts – from the real world
False facts – not studied in experiment and
not true in the real world
 Fan effect occurs with all three fact types
Interference from Preexisting
Facts
Interference vs Decay
 Less forgetting during sleep than when
awake.


Occurs because material is retained better
when learned at night.
Night is period of highest arousal.
 Forgetting functions may reflect
interference from unknown sources.
 Decay theories do not specify any
mechanism for decay.
Effects of Redundancy
 Interference occurs only when learning
multiple memories that have no
relationship to each other.
 Bradshaw & Anderson – compared
relevant and irrelevant fact learning:


Irrelevant facts interfere.
Relevant facts aid memory compared to
single fact learning.
Relevant vs Irrelevant
Information
Immediate Recall
%
Recall at 1 week
%
Single fact
92
62
Irrelevant facts
80
45
Relevant facts
94
73
Retrieval and Inference
 Much of memory is inference at the time
of recall – not actual retrieval of facts.
 Bransford et al. -- inference can lead to
incorrect recall:


Turtles resting on or beside a log and a
fish swam beneath them.
Subjects were most confused by test
sentences whose meaning was implied by
the studied sentences (“beneath it”).
Carol Harris’s Story
Carol Harris was a problem child from
birth. She was wild, stubborn, and
violent. By the time Carol turned eight,
she was still unmanageable. Her parents
were very concerned about her mental
health. There was no good institution for
her problem in her state. Her parents
finally decided to take some action. They
hired a private teacher for Carol.
Inference-Based Intrusions
 Sulin & Dooling – subjects add details
not present during learning:



Carol Harris vs Helen Keller
“She was deaf, dumb and blind.”
5% Carol Harris but 50% Helen Keller
subjects falsely recognized the sentence.
 Inferences are made at test-time.

More inferential errors occur with delay.
Plausible Retrieval
 Reder – much of recall is plausible
inference not actual recall.


Darth Vader inferred to be evil, not
remembered to be evil.
Heir to hamburger chain story – subjects
asked to recall exact details and make
plausible inferences.
 After a delay, plausible inference is
faster and does not decay as much as
exact memory, with no fan effect.
Plausible Retrieval after a
Delay
Inference and Elaboration
 Elaboration leads to more inferences.

Information added as a “theme” to a story
results in better recall of studied material
and more inferences (Nancy & the doctor).
 Intruded inferences are not necessarily
“errors” but help guide our thinking and
behavior.
 Listerine court case – false inferences,
not just false statements, not permitted.
Memory Errors
 When exact memory is needed,
inferences and reconstructive processes
can be misleading.
 Loftus -- additional details and
suggestion can change what is recalled.
 John Dean’s recall vs what Nixon
recorded – gist was right but not details.
 False memory syndrome – memories
that never happened can be “planted.”
Deese-Roediger-McDermott
“False Memory” Paradigm
 Participants study lists of words:


Thread, pin, eye, sewing, sharp, point, prick, thimble,
haystack, thorn, hurt, injection, syringe, cloth, knitting
Bed, rest, awake, tired, dream, wake, snooze,
blanket, doze, slumber, snore, nap, peace, yawn,
drowsy
 Later shown 3 types of words:


True (sewing, awake), False (needle, sleep), New
(door, candy)
Participants accept 88% of true words & 12% of new,
but also accept 80% of false words.
False Memories and the Brain
The parahippocampus
retains the sensory
experience of seeing
the word. The
hippocampus extracts
and stores the meaning.
Context Effects
 Recall is better if the physical context
during learning is also present during
testing.


Experimenter clothing, setting.
Under water.
 Eich suggests that context effects
depend on integrating context and the
material to be learned.
Context Effects
Mood Congruence
 Bower et al. – hypnotized subjects and
induced positive or negative mood.


Recall better if hypnotized into the same
mood during testing as during learning.
Again, the effect may depend upon
integration of mood with material learned.
 Mood congruence – easier to remember
memories congruent with the current
mood.
Mood Congruence
State-Dependence
 Material is easier to recall if people
return to the same emotional and
physical state as during learning.



Drinking – some state dependence
together with overall debilitating effect on
memory.
Marijuana and tobacco.
Caffeine.
 Studying when not intoxicated is better.
State-Dependent Learning
Cigarette
Marijuana
Average
Cigarette
25
20
23
Marijuana
12
23
18
Encoding Specificity
 The other items presented during
learning provide a context too.

Presentation of cues in as close to the
original learning context aids recall.
 Encoding specificity principle:

The probability of recalling an item
depends on the similarity of its encoding at
test to its original encoding at study.
Test of Encoding Specificity
 Watkins & Tulving:




Study pairs of words
Generate associates for words & indicate
which were among studied words.
Cued with first word of pair.
61% recall in cued task, <54% in associate
recognition task.
 Recognition generally produces higher
scores so result should have been the
opposite of what occurred.
Amnesia
 Studies of amnesics tell us how memory
is organized in the brain.
 Amnesia occurs with damage to the
hippocampus (and some other areas).
 Kinds of amnesia:


Korsakoff syndrome
Retrograde vs anterograde amnesia
 Patient H.M.
Types of Amnesia
RA: Retrograde Amnesia
AA: Anterograde Amnesia
What is Spared in Amnesia?
 Memory for facts, knowledge of
meanings of words, language.
 Memory for how to do things (e.g., play
the piano, tie shoes), skills.
 Priming
 Incidental learning – memory for
experience that was not consciously
attended to.
 Working memory – short term memory.
What is Affected by Amnesia?
 Episodic memory – memory for the
details and experiences of one’s own
life.
 Learning and recall of new material -anterograde amnesia

Because conscious learning starts out as
an episodic experience.
Implicit vs. Explicit Memory
 Explicit memory – knowledge we can
consciously recall.
 Implicit memory – knowledge we cannot
recall but which aids performance on a
task.
 Amnesics can do a word-completion task
but not recall learned words.
 Normal subjects also show an explicitimplicit dissociation – Jacoby’s study.
Recall vs Stem Completion
Jacoby’s Study of Normal
Subjects
 Three conditions were presented:



Study a word alone
Study a word with its antonym
Generate the word as an antonym
 Two tests were used:


Recognition task (old/new items) – explicit
Perceptual ID task (see for 40 ms) –
implicit
 Opposite effects for the two tasks.
Recognition vs Perceptual ID
Procedural Memory
 Procedural memory can be for skills, but
also for doing cognitive tasks.
 Berry & Broadbent – control output of
hypothetical sugar factory by changing
size of workforce (computer simulation):


Non-obvious formula involved.
After 60 trials subjects were good at task
but could not state the rule involved.
 Amnesics can learn to do this too.
Squire’s Varieties of Memory