Neuropsychology, Delusions and Modularity:

1
Neuropsychology, Delusions and Modularity The
Curious Problem of Belief
www.neuro.spc.org/psypag
Vaughan Bell vaughan_at_backspace.org
Hadyn Ellis Peter Halligan
2
Outline

• Delusions and Belief.
• Cognitive Neuropsychology and Current Conceptual
  Tools.
• Problems with modular approaches to belief.
• Approaches for crossing the divide.

3
Notable Delusions

• Shot by John Hinckley, who believed that this
  would cause Jodie Foster fall in love with him.

4
Notable Delusions
Philip K Dicks writing often drew on delusional
experiences.
5
Clinical Definition

• Delusions, as defined by psychiatrists are
• fixed false beliefs based on an incorrect
  inference about external reality that is firmly
  sustained despite what almost everyone else
  believes and despite what constitutes
  incontrovertible and obvious proof or evidence to
  the contrary
• (DSM-IV-TR)

6
Delusion / Belief Relationship

• Despite the fact that almost every one of these
  points can be argued with (Spitzer, 1990 David,
  1999)
• Delusions are, by and large, considered as
  beliefs (albeit pathological ones).
• Therefore an attempt to understand the
  neuropsychology of delusions, needs to be able to
  tackle the neuropsychology of belief.
• As is the explicit objective of Cognitive
  Neuropsychiatry in this area.

7
Traditional CN Approach

• Traditionally, cognitive neuropsychology has
  relied on several assumptions to allow the study
  of pathology to uncover normal functioning.
• Modularity is one such assumption.
• It assumes that fundamental cognitive process are
  independent and encapsulated within the brain.
• With these assumptions we can uncover previously
  unknown modules by studying differences in
  behaviour after pathology.

8
The Belief Stumbling Block

• Fodor (1983) the originator of the modularity
  hypothesis says that belief is a central
  process.
• That is, is does not exist as a modular entity
  within the brain.
• Indeed the relation of brain function to belief
  is still particularly controversial.
• For example, Churchland (1999) argues that
  belief is an outdated concept.
• And we will not be successful in determining its
  relation to brain function.

9
The Belief Stumbling Block

• This would suggest methods that rely on
  modularity assumptions will less useful in
  understanding belief.
• e.g. The lesion method.
• Which seems to suggest a bleak future for
  Cognitive Neuropsychiatry in this area
• and its researchers who wish to study delusions
  as a way of understanding normal belief
  processes.

10
Possible Solutions

• Luckily, a similar problem has been encountered
  before.
• Burgess (1997) has argued that the executive
  system is also a central process.
• And has identified particular techniques for a
  successful cognitive neuropsychological approach
• ..which could be used in cognitive
  neuropsychiatry to understand the significance of
  delusions.

11
Heterogeneous Samples

• Heterogeneous patient samples can be used.
• Rather than grouping similar patients, by either
  syndrome (e.g. schizophrenia) or by delusion type
  (e.g. persecutory)
• we can select a symptom (delusion) and aim to
  include multiple aetiologies in the group.
• Common behavioural effects can then be extracted
  post-hoc
• ..and related to the common psychopathology (in
  this case pathologies of belief).

12
Holistic Belief Models

• Many theories talk about the web of belief but
  at the moment this is still a black box.
• Coherence Theory (Thagard, 2000) gives one
  possible way of modelling this.
• Element (beliefs) have constraints between them
  to either encourage connected beliefs to be
  accepted or rejected.
• Elvis lives vs Elvis is buried at Graceland
• We aim for maximum coherence in our belief set.

13
Coherence Theory Approach

• Computing maximum coherence is a computationally
  intractable problem.
• But Hopfield style interconnected neural nets are
  very good at creating best fit solutions.
• This allows for dynamic and holistic modelling of
  belief states.
• As well as the ability to test theories of
  pathology on the network, leading to testable
  predictions.

14
Conclusion

• Cognitive neuropsychiatry as applied to belief
  presents some unique challenges.
• The neuropsychological approaches derived from
  theories of modularity may not be as useful.
• Work from similar challenges in executive system
  research may be fruitful.
• Novel approaches (such as coherence theory) may
  be needed to understand wider issues in the
  neuropsychology of belief and pathologies of
  belief.