Transcript Lecture 4

Key points for lecture 4
• Forgetting: good or bad?
• How does the Deese / Roediger / McDermott (DRM) paradigm
work?
• Which factors increase or decrease false memory?
• What are the actual underlying (cognitive) causes of false memory?
• Reisberg, Chapter 7.
When Memories Go Wrong
• What happens when your memory of an event does
not correspond to what actually happened?
– In what ways can our decisions get warped by an
inaccurate memory?
– Are we always aware when this happens?
Forgetting
• Is Normal!
• And desirable!
• The Case of “S” (Luria, 1968)
The Case of “S”
• “S” did not benefit a great deal from having
a ‘perfect’ memory.
– Impaired ability to abstract general knowledge
from his experiences.
– Related to his inability to forget specific details
of each event?
– Almost the opposite of Varga-Khadem’s
amnesic children.
Episodic vs. Semantic Memory
• Baddeley’s Metaphor
• Our general knowledge
is represented in a
distinct ‘semantic’
Memory
Episodic & Semantic Components of
Autobiographical Memory
• Parker et al. (2006), Neurocase 12: pages 3539. (pdf available via my webpages)
• Case A.J. : ‘…highly superior semantic
autobiographical memory’
Sources of Error in Normal Memory
• Forgetting.
– A natural feature of our memory?
• Recollection and familiarity may have to
trade off against one another all the time.
– How might their interaction distort our memory
of the past, and mislead our judgement?
The Weight of Eyewitness Evidence
• An estimated 77,000 people annually in the
USA are charged solely on the basis of eye
witness evidence.
• Around three quarters of English cases
result in conviction due to eye witness
testimony (of which half were based on a
single eye witness).
Introducing Distortions into Memory
• Force subjects to experience very similar
kinds of episodes, which become hard to
discriminate from one another
• Manipulating the familiarity of retrieval
cues
The Deese (1959) Recall Task
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
SOUR
CANDY
SUGAR
BITTER
GOOD
TASTE
TOOTH
NICE
HONEY
SODA
CHOCOLATE
HEART
CAKE
TART
PIE
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
QUEEN
ENGLAND
CROWN
PRINCE
GEORGE
DICTATOR
PALACE
THRONE
CHESS
RULE
SUBJECTS
MONARCH
ROYAL
LEADER
REIGN
The Deese (1959) Recall Task
• Deese constructed his lists using word
association norms.
• Each item in a list is a strong associate of a
particular TARGET word.
• Deese found high levels of recall intrusions
by these unpresented TARGET items.
Roediger and McDermott (1995)
• Modified and extended Deese’s basic result.
– Employing recall and recognition tasks
– Use of the Remember / Know (R/K) procedure.
The Remember / Know Procedure
• Ask subjects to report on their experiences
while recognising.
– Do they ‘Remember’ any episodic details?
– Or do they just ‘know’ the information was
encountered at study?
Recognition Test List
–
–
–
–
–
–
–
–
–
–
–
–
–
PLACE
SWEET
TABLE
PARTY
GENERAL
MEMORY
CONSENSUS
KING
COMPUTER
TREE
FERRET
BURGLAR
BOTTLE
Roediger and McDermott (1995)
• Percent Recognition
90
80
70
60
50
TRUE
FALSE
40
30
20
10
0
OVERALL
REMEMBER
KNOW
Some Factors that increase or
decrease DRM False Memory
• Increase: the number of associates presented for
study
• Increase: the strength of association between study
list items and their TARGET
• Decrease: (in young people) multiple study-test
cycle.
• Decrease: the ‘distinctiveness heuristic’
‘Distinctiveness Heuristic’
• Two study conditions
– Words from the DRM lists
– Words from the DRM lists paired with a picture
• False recognition was almost absent when
words had been paired with pictures
• The ability to recollect picture information
was ‘diagnostic’ for studied items.
A triple whammy!
3 Reasons for ‘DRM’ False Memory
(1) Implicit associative responses
– subjects themselves generate the target items while
studying each list.
– Then experience ‘source confusions’ at test
(2) Familiarity of ‘lure’ items
– But what about the ‘Remember’ responses?
(3) A loss of encoding specificity
The puzzle raised by false memory
• Within the consensus view, how is it possible to recollect
events that never took place?
• That is, what might cause
Source errors?
Familiarity-based confusions?
Loss of encoding specificity?
CMF Explanations for DRM False Memory
• The hippocampal formation



Pattern separation failure at encoding
Pattern completion failures at retrieval
Therefore: source errors, & loss of encoding specificity
• The frontal lobes


Strategic control over memory
Failure to adequately focus on cues and/or monitor retrieval
• The entire ‘association’ neocortex

Represents very similar content across a succession of episodes
Summary
• Judgements are most accurate when they
are made on the basis of information whose
source has been recollected.
• But if retrieval instructions allow it,
judgements may be based, by default, upon
potentially less accurate familiarity.
–