How automatic is familiarity really ERP evidence from recognition memory experiments

1
How automatic is familiarity really? ERP
evidence from recognition memory experiments

• Ullrich Ecker
• Saarland University, Germany
• University of Western Australia
• Hubert D Zimmer
• Saarland University, Germany

2
ERwhat?

• EPC 2008 programme count
• ERP/ Event-related potential 1 (this talk)
• EEG 0
• fMRI 1
• neuro- 1
• Neuron 1
• Brain 1
• Model 497
• Categorization 312
• Decision 253

3
ERP introduction

• A quick EEG/ERP introduction...

4
ERP introduction
5
Electrogenesis
What's EEG?

• Brain electrical activity at the scalp
• ? Voltage over time

6
Electrogenesis
Open-field theory

• Voltage is generated by
• cell clusters, that are
• active synchronously and compose an
• open field

7
The EEG
The extended 10-20-system
3 x 3 grid of ROIs
8
The EEG
Localisation problem

• Bad localisation
• The brain is a volume conductor, so any given
  activity may be generated distally (e.g.,
  subcortical far field potentials)

9
EEG ? ERP
Signal vs. noise

• EEG is composed of
• event-related (evoked) activity (the signal)
  and
• spontaneous activity plus artefacts (the noise)

• To get rid of noise
• filtering (e.g., 0.5 Hz 20 Hz)
• ocular correction (algorithm)
• artefact correction (exclusion of trials)

10
EEG ? ERP
Signal extraction by way of averaging
Event-related averaging (condition-wise)
11
EEG ? ERP
Signal extraction by way of averaging
12
ERP
Components

• What's a component and what does that mean?
• physiological approach
• Where is the source of the signal (source
  analysis)?
• psychological approach
• What (cognitive) process is reflected in the
  component?

13
Method
ERP old-new effects recognition memory
mid-frontal
µV
mid-frontal old-new effect 300 500
ms familiarity
0
5
CRs Hits
10
0
500
ms
left-posterior
µV
0
left-parietal old-new effect 500 700
ms recollection
5
10
0
500
ms
14
Method
ERP old-new effects recognition memory
mid-frontal

• Mid-frontal old-new effect
• 300 500 ms
• Familiarity
• Not LOP sensitive
• Larger for K vs R responses
• Sensitive to criterion changes
• Covaries with recognition confidence
• Left-parietal old-new effect
• 500 700 ms
• Recollection
• - LOP sensitive
• - Larger for R vs K responses
• Less sensitive to criterion changes
• Confidence effects unclear
• Specifically affected by certain drugs

µV
0
5
CRs Hits
10
0
500
ms
left-posterior
µV
0
5
10
0
500
ms
15
Common opinion
Claims about familiarity

• CLAIM 1 Familiarity is a relatively automatic
  process.
• EVIDENCE
• Familiarity (vs. recollection) is fast
• Divided attention during both encoding and
  retrieval has less detrimental effects on
  familiarity (vs. recollection e.g. Troyer
  Craik, 2000, CanJEP)
• Frontal lobe pathology (age, lesions) has small
  effects on familiarity (vs. recollection
  Janowsky et al., 1989, Neuropsychologia
  Yonelinas, 2002, JML)

16
Common opinion
Claims about familiarity

• Contrary to some claims
• (familiarity conceptual priming see Paller et
  al., 2007, TICS)
• CLAIM 2 Familiarity is a perceptually specific
  process.
• EVIDENCE
• Perceptual study-test manipulations impair
  recognition performance in amnesics (Srinivas
  Verfaellie, 2000, JML)
• Mid-frontal ERP old-new effect is sensitive to
  perceptual study-test overlap
• Schloerscheidt Rugg, 2004, Neuropsychologia
• Groh-Bordin, Zimmer Mecklinger, 2005, Cog Brain
  Res
• Groh-Bordin, Zimmer, Ecker, 2006,
  Psychophysiology
• Ecker, Zimmer, Groh-Bordin, 2007, Mem Cognition
• Ecker, Zimmer, Groh-Bordin, 2007, Brain Res
• See also
• Rugg Curran, 2007, TICS
• Grove Wilding, in press, J Cog Neurosci

17
Example
Design
STUDY TEST
? accept (Old/Same)
? accept (Old/Different)
? reject (New)
intentional study ... inclusion task (feature
irrelevant)
18
Example
Effects associated with changing task-irrelevant
feature from study to test
Performance costs
Perceptual specificity effects
Graded ERP familiarity signal
New (CRs) Different (hits) Same (hits)
(Ecker et al., 2007, Mem Cognition)
19
Example study
Graded ERP familiarity signal reflecting
perceptual study-test similarity
(Ecker et al., 2007, Brain Res)
20
Common opinion
Claims about familiarity

• CLAIM 1 Familiarity is a relatively automatic
  process.
• EVIDENCE
• Familiarity (vs. recollection) is fast
• Divided attention during both encoding and
  retrieval has less detrimental effects on
  familiarity (vs. recollection e.g. Troyer
  Craik, 2000, CanJEP)
• Frontal lobe pathology (age, lesions) has small
  effects on familiarity (vs. recollection
  Yonelinas, 2002, JML)
• Mid-frontal perceptual specificity effects do not
  depend on task-relevance of the feature

21
Common opinion
Claims about familiarity

• Taken together
• Familiarity is considered an automatically
  supplied graded signal
• Coding for both conceptual and perceptual
  features of objects.
• But

22
How automatic is familiarity really?
Boundary conditions

• There is evidence that a familiarity signal is
  only elicited when subjects are in a retrieval
  mode
• Tonically maintained cognitive state biasing the
  system towards treating external events as
  retrieval cues
• (Rugg Wilding, 2000, J Cog Neurosci Tulving,
  1983)
• Memory-related old-new effects strongly affected
  by retrieval intention (Düzel et al., 1999, PNAS)
• No mid-frontal old-new effect in implicit tasks
• (Groh-Bordin et al., 2005, Cog Brain Res)

23
How automatic is familiarity really?
Boundary conditions

• What role does retrieval orientation play?
• Highly specific form of processing applied to a
  retrieval cue
• Tonically maintained state
• (Herron Wilding, 2004, NeuroImage Rugg
  Wilding, 2000, TICS)
• Contribution of retrieval orientation to the
  moulding of familiarity/recollection and the
  associated ERP old-new effects is unclear (see
  Hornberger, Morcom, Rugg, 2004, J Cog Neurosci)

24
Questions
Retrieval orientation and familiarity

• Are subjects able to adapt flexibly to differing
  retrieval orientation demands?
• (Johnson Rugg, 2006, Brain Res Herron
  Wilding, 2004, NeuroImage vs. Koutstaal, 2004,
  PBR Herron Wilding, 2006, NeuroImage)
• Can we corroborate that familiarity is based on
  conceptual and perceptual processing?
• Does familiarity and the ratio of
  perceptual/conceptual processing associated with
  it depend on retrieval orientation?

25
Design
(Ecker Zimmer, under review)
Study (nat/art) Test Test Cue (1.5 s
pre S) Test Condition Correct response
category exemplar
exemplar category exemplar
category
Same Different New New
Same Different Old
New New New
Old Old
26
Hypotheses
Behavioural
(Ecker Zimmer, under review)

• Overall, subjects should be able to flexibly
  adjust their retrieval orientation on a
  trial-by-trial basis
• Different items call for differential response
  across Category/General and Exemplar/Specific
  conditions
• Replication of Koutstaal (2004, PBR) pattern
• Memory performance should depend on
• Study-test-overlap (perceptual and conceptual)
• Response conflict
• In particular, even in the Category/General
  condition Same gt Different

27
Results
Behavioural data
(Ecker Zimmer, under review)


28
Results
Behavioural data
(Ecker Zimmer, under review)


29
Hypotheses
ERP
(Ecker Zimmer, under review)

• Familiarity is perceptually specific
• ? Mid-frontal old-new effect Same gt Different
  (perceptual specificity effect)
• If familiarity is strategically modulated, the
  retrieval orientation manipulation should impact
  differentially on this perceptual specificity
  effect
• Category/General focus conceptual processing
  should be more important
• Exemplar/Specific focus item processing biased
  towards perceptual appraisal
• ? Larger perceptual specificity effect in the
  Exemplar/Specific condition
• ? mid-frontal old-new effect for Different items
  only in Category/General condition

30
Results
ERP data Category/General condition
(Ecker Zimmer, under review)
31
Results
ERP data Exemplar/Specific condition
(Ecker Zimmer, under review)
32
Questions
Retrieval orientation and familiarity

• Are subjects able to adapt flexibly to differing
  retrieval orientation demands?
• Can we corroborate that familiarity is based on
  conceptual and perceptual processing?
• Does familiarity and the ratio of
  perceptual/conceptual processing associated with
  it depend on retrieval orientation?

33
Answers
Retrieval orientation and familiarity

• Are subjects able to adapt flexibly to differing
  retrieval orientation demands?
• Yes. Retrieval orientation is flexible (no
  tonically maintained state)
• Can we corroborate that familiarity is based on
  conceptual and perceptual processing?
• Yes. Familiarity can be conceptually and
  perceptually specific (conveys perceptual details
  of study episode)
• Does familiarity and the ratio of
  perceptual/conceptual processing associated with
  it depend on retrieval orientation?
• Yes. Familiarity is influenced by top-down
  modulation of retrieval orientation
• The ratio of perceptual/conceptual processing
  depends on task demands (e.g., no mid-frontal
  perceptual specificity effects if study is
  conceptual)
• ? Thus, familiarity is strategically modulated
  and not fully automatic
• (hence it is not equivalent to conceptual
  priming)

34
Thank you!
Hubert D. Zimmer Steve Lewandowsky
35
Thank you!
Cognitive Psychology Building Saarland
University, Germany
Cognitive Psychology Building University of
Western Australia
36
Discussion

• Dual process theories assume recognition is
  based on recollection and familiarity, with only
  recollection providing knowledge about study
  details.
• Dual process models were able to account for the
  data only when they incorporated continuous
  evidence a familiarity signal sensitive to
  details.
• (from Heathcote, Raymond, Dunn, 2006, JML)

37
Discussion
No mid-frontal perceptual specificity effects if
study is conceptual
Consistent with classic dual-process theory
details conveyed in recollection Familiarity
rather conceptual
Verbal/Conceptual study
(Ecker et al., submitted)