Category Learning and Dopaminergic Personality Traits: Evidence for a Hippocampal Contribution to So

1
Category Learning and Dopaminergic Personality
Traits Evidence for a Hippocampal Contribution
to Some Task Variants?

• Alan Pickering
• Department of Psychology
• a.pickering_at_gold.ac.uk

2
Collaborators

• Various students from Psychology Departments
  within the University of London
• Harry Pidd (Goldsmiths College)
• Rozmin Halari (Institute of Psychiatry)
• Lucy Schomberg Luke Jones
• (St Georges Hospital Medical School)

3
Evidence for Influence of DA on (Category)
Learning

• Neuropsychological
• Studies with patients with damage to DA-rich
  striatal brain regions are impaired on some
  category learning tasks
• Neurophysiological
• Studies suggest synaptic plasticity in striatal
  brain regions require appropriately timed DA
  signal

4
Category Learning in Parkinsons Disease

• Weather task Knowlton et al, 1996

5
Category Learning in Parkinsons Disease

• Main Findings Knowlton et al, 1996

6
What Processes in Learning Tasks Might be
DA-Related?

• Two main candidates are widely proposed

• Appetitive reinforcement and motivation
• DA activity has been viewed as provided a
  reinforcement signal which is required for
  synaptic plasticity (3-factor learning rule)

7
DA-Related Processes (cont)

• Input Modulation
• Schultz (1998) DA lets only the strongest
  cortical influences pass to striatal neurons,
  whereas weaker inputs become ineffective
• Floresco et al (2001) DA receptor activity
  serves to strengthen salient inputs while
  inhibiting weaker ones

8
Logic of The Present Studies

• Certain biologically-based personality traits in
  the healthy population reflect variation in
  dopaminergic neurotransmission
• Therefore, these traits should correlate with
  performance on DA-sensitive category learning
  tasks (e.g. those involving reinforcement or
  input modulation)

9
Which Personality Traits Are DA-Related?

• Extraversion-Introversion (Depue, Rammsayer)
• Example measures EPQ-E Introvertive Anhedonia
  (IntAnh)
• Impulsive Antisocial Sensation Seeking ImpASS
  (Gray, Cloninger, Zuckerman, Pickering)
• Example measures EPQ-P Novelty Seeking
  Sensation Seeking Scale
• Positive Schizotypy (many)
• Example measures Unusual Experiences

10
Example Questionnaire Items

• ImpASS Measure EPQ-P (25 items)
• -Have people said that you sometimes act
• too rashly?
• -Should people always respect the law?
• -Would you take drugs which may have
• strange or dangerous effects?
• Positive Schizotypy Measure Unusual
  Experiences (30 items)
• -I have felt that I have special, almost
• magical powers
• -Do you ever feel that your thoughts dont
• belong to you
• -Sometimes my thoughts are as real as
• actual events in my life

11
What Evidence Implicates DA in Personality Traits?

• Direct evidence for extraversion and ImpASS
• -Traits correlate with psycho-physiological
  response to DA drugs
• -Traits correlate with DA-binding in
  neuroimaging studies
• -Traits correlate with DA receptor gene
  variations
• No evidence for positive schizotypy but argument
  by analogy with schizophrenia schizotypal
  personality disorder

12
DA and ImpASS Sample Evidence

• Gray, Pickering Gray (1994) SPET D2-binding in
  basal ganglia and EPQ-P

13
Category Learning ImpASS Scores Early Studies

• Ball Zuckerman (1990) positive correlation
  between Sensation Seeking scores and learning of
  a concept formation task
• Pickering (2 studies) positive correlations for
  Novelty Seeking and EPQ-P scores with learning
  of a filtration category learning task
• Suggestive but uninformative about what
  DA-related processes might be involved

14
My Personality and Category Learning Research
Programme

• Series of studies which attempt to relate
  specific task processes to specific (DA-related)
  traits
• Model results using biologically-constrained
  neural networks, varying DA parameters to try to
  model trait variation
• Results show dissociable relationships across
  traits and tasks

15
Overview of Results

• Anx anxiety (control trait)

16
Aims of Current Talk

• To illustrate some of the results
• Reinforcement Study suggesting extraversion
  and positive reinforcement link
• Focus on the possible links between ImpASS traits
  (esp. the EPQ-P measure) and hippocampal system
  processes during category learning and related
  tasks

17
Reinforcement Study

• Pickering and Halari
• Within-Ss design using 2 equivalent probabilistic
  category learning tasks Weather task and a
  Symptoms-Disease task
• Training
• Task 1 enhanced reinforcement
• 0.10 per correct response
• Task 2 paired-associate training
• meant no reinforcement
• Testing
• Categorise each stimulus without reinforcement

18
Details

• 40 healthy male participants, mostly students
• Personality Measures
• Extraversion Introvertive
• Anhedonia (IntAnh)
• ImpAss EPQ-P
• Schizotypy Unusual Experinces
• (UnEx)
• Dependent Variable Accuracy of responses during
  test

19
Results Correlations

• RF reinforcement task score
• PA paired-associate task score

20
Summary of Findings

• DA-related personality traits again correlated
  with performance on DA-sensitive category
  learning tasks
• Extraversion-introversion (but not ImpAss or
  schizotypal traits) correlated with learning
  under appetitive reinforcement component
• ImpAss correlated with learning under
  paired-associate training

21
Interpretation of ImpASS Result

• Could the paired-associate training encourage a
  greater dependency on hippocampally-mediated
  explicit remembering?
• Do ImpASS traits relate to functioning of
  hippocampal system?
• Cons Probabilistic training
• Pros Subsequent paper by Poldrack et al
  (Nature, 2001)

22
Poldrack et al 2001

• fMRI study in healthy volunteers with weather
  task under standard feedback (FB) and also paired
  associate (PA) training
• Medial temporal lobe (MTL) activity higher for PA
  than for FB task
• Reverse was true for caudate nucleus (CN)
  activity
• Negative correlation between MTL and CN
  activation in both tasks

23
ImpASS Traits and Hippocampal Functioning

• Little existing evidence
• Some evidence from human Latent Inhibition tasks
• We carried out several studies to test the link
  further
• - latent inhibition study
• - paired-associate learning
• - associative mismatch detection

24
ImpASS Traits and Latent Inhibition (LI)

• LI tasks in animals are sensitive to hippocampal
  lesions
• Human LI analogues studied in relation to
  positive schizotypy and ImpASS
• Pickering Gray (2001) review shows LI
  associated with either type of trait tested
  separately
• High trait scores and hippocampal lesions affect
  performance in same direction

25
LI Study Pickering Pidd

• Studied LI in relation to measures of both
  positive schizotypy and ImpASS
• Used standard human LI task (learning association
  between a specific sound on a tape and a counter
  increment after preexposure to sound without
  increment)
• Tested 86 subjects mostly students

26
Pickering Pidd Details

• Trait Measures
• Positive Schizotypy
• - Unusual Experinces (UnEx)
• ImpASS
• - Impulsive Nonconformity (ImpNon)
• - Sensation Seeking Scale (SSS)
• Nonparametric correlations between LI task
  performance (trials to criterion) and trait
  measures

27
Pickering Pidd Results
After partialling out UnEx
Spearmans rho (all plt0.001)
R20.29
28
Conclusions

• ImpASS and positive schizotypy trait measures
  make independent contributions to learning in the
  preexposed condition of a human LI task
• In both cases, high trait scores facilitate
  learning (reduce LI) the same effect as
  hippocampal lesions
• Suggests ImpASS trait scores may reflect
  individual differences in hippocampal system
  functioning

29
ImpASS and Paired Associate Learning I

• Pickering and Schomberg
• Unrelated verbal paired-associates (e.g.
  SOIL-MILE SIDE-BRAVE) were used
• This is the quintessential explicit memory task
  sensitive to hippocampal lesions
• 40 healthy subjects (students)
• Extraversion (Ext), ImpASS
• (EPQ-P), and positive schizotypy (UnEx) were
  measured

30
ImpASS and Paired Associate Learning II

• 12 word pairs (A-B) used
• 3 study-test learning trials
• Test cued recall for B using A as cue
• 1 unexpected 10-min delayed cued recall test
  trial
• DVsNumber correct on each test (NC1, NC2, NC3,
  NCD)
• Measured IQ subtest performance for each subject
  (WAIS-III Matrices)

31
ImpASS and Paired Associate Learning Results

• denotes correlation with EPQ-P after
• partialling out IQ

32
Interpretation Conclusions

• ImpASS traits appear to be positively associated
  with performance on hippocampal-sensitive tasks
• Given ImpASS traits also correlate with
  performance on standard and paired-associate
  category learning (CL) tasks (see above), this
  suggests a possible hippocampal contribution to
  CL tasks

33
Key Questions Raised

• What specific processes in explicit remembering
  might be linked to ImpASS traits?
• What about links between ImpASS traits and DA
  neurotransmission?
• What evidence exists on role of DA activity in
  hippocampal system?

34
Dopamine and Hippocampus

• Work by Otmakhova Lisman has demonstrated that
  DA reduces transmission in the perforant path
  projection to CA1 but not in the CA3-CA1 pathway
  (Schaffer collaterals)
• This modulatory DA effect depended on both
  D1-like and D2-like receptors

35
Otmakhova Lisman (1999)

• J Neurosci., 19, 1437-1445

36
Otmakhova-Lisman Model I
CA3 outputs prediction for input at time t1
based on input at time t
Dopamine suppresses this input specifically
Input at time t1
Input at time t
CA3
CA1
CA1 compares predicted input with actual input
Match or Mismatch
37
Otmakhova-Lisman Model II

• Many similar models (Vinogradova, 1984 Levy,
  1989 Gray McNaughton, 2000) but how do they
  account for explicit memory functions?
• Underactivity of D1/D2 DA receptors would make
  hippocampus oversensitive to mismatch
• High ImpASS scoring subjects have hypofunctional
  D2 receptors
• Prediction High ImpASS subjects will have
  higher mismatch sensitivity than low ImpASS

38
Associative Mismatch Study

• Pickering and Jones
• Choice reaction time (RT) task
• Warning stimuli for RT trials manipulated to
  create regular sequence which subjects learn
  incidentally
• Observe effect on RT task during critical trials
  (e.g. when mismatching warning stimulus sequence
  is used)

39

Standard Trial Sequence
A
40

Associative Mismatch Sequence
L
41
Choice RT Task Measures

• Mean RT and errors on
• critical trials (e.g. associative mismatch, AM)
• compared with
• matched control trials (i.e. immediately
  preceding standard trials)
• For correlations use difference
• measures
• (critical RT matched control RT)

42
AM Study Details

• Investigated effects of AM and secondary
  reinforcers (SRs), of 4 different types (aversive
  and appetitive)
• 40 healthy volunteers (male medical students)
• Used preconditioning procedure to create
  secondary reinforcers
• 350 choice RT trials (1-50 practice), including 6
  AM and 4x6 SR trials

43
AM Study Questionnaires Used

• Extraversion-Introversion
• - EPQ-E
• - Introvertive Anhedonia (IntAnh)
• ImpASS
• - EPQ-P
• - Novelty Seeking Scale NS
• Anxiety
• various (results not reported here)

44
Mismatch Study Overall RTs
Trials with various reinforcer types
AM associative mismatch trials
45
Mismatch Study Results

• Mean (s.d.) mismatch RT difference 40.1 57.8 ms

Outliers So we used nonparametric correlations
and parametric correlations on Winsorized data
46
Mismatch Study Results

• Correlations with mean RT difference
• on Mismatch Trials

47
Summary of Mismatch Results

• As predicted, ImpAss traits are positively
  associated with sensitivity to associative
  mismatch
• Other data from our lab (e.g. Gray, Pickering, et
  al, 2002) on rule learning tasks is consistent
  with these observations
• High ImpASS subjects are more affected by
  unpredicted rule shifts than low ImpASS

48
General Conclusions

• These data reinforce the view that dopaminergic
  activity plays multiple roles in category
  learning (CL)
• A significant association between CL performance
  and some personality traits may be a signature
  for involvement of particular psychological
  processes
• ImpASS traits may influence CL particularly when
  the task engages hippocampal mismatch detection
  mechanisms