Tuesday, 11 March 2014

Explaining Delusional Belief: The Two-Factor Account

Max Coltheart
I'm Max Coltheart, Emeritus Professor of Cognitive Science at the Centre for Cognition and its Disorders and Department of Cognitive Science, at Macquarie University. I work in cognitive neuropsychology (especially developmental dyslexia) and cognitive neuropsychiatry (especially delusional belief). I am also interested in the use of functional neuroimaging to attempt to test cognitive theories. I did a joint undergraduate degree in psychology and philosophy at the University of Sydney and have worked with the philosophers Martin Davies, John Sutton and Peter Menzies.

Since the late 1990s my colleague Robyn Langdon and I, initially in collaboration with Martin Davies and the clinical neuropsychologist Nora Breen, and later in collaboration with various others especially Ryan McKay, have been developing a cognitive-level theory of the genesis of delusions which we call the two-factor theory. Our view is that scientific understanding of any delusional condition has been achieved if the answers to two questions has been discovered. Question 1 is: what caused the idea that is the content of the delusion to first come to the deluded person’s mind? Question 2 is: what caused this idea – a candidate for belief - to be accepted as a belief, rather than rejected, which is what ought to have happened because there is so much evidence against it?

Our theory stems from ideas about delusion put forward by William James in 1890 and Brendan Maher in 1974. Both proposed that delusional beliefs result from the use of intact normal reasoning processes in an attempt to develop hypotheses to explain abnormal experiences that the patient is having. These abnormal experiences correspond to factor 1 in our theory, and we agree with James and Maher that it is the application of normal processes of inference (we suggest it is abductive inference that is used here) that yield the hypotheses that can become delusional beliefs.

Where we depart from James and Maher is in our view that the hypotheses abductively inferred from the abnormal experiences as attempted explanations of these experiences will normally be evaluated, and then rejected because of the strength of the evidence against them. That would prevent these candidates for belief from becoming actual (delusional) beliefs. Only if a second factor is also present, an impairment of the belief evaluation system, will patients fail to reject the hypotheses suggested by the abnormal experiences. So only when this second factor is present will delusional beliefs arise. (See Coltheart, Menzies, and Sutton 2009).

Consider the Capgras delusion: the belief that someone emotionally close to you, often a spouse, has been replaced by a complete stranger. Haydn Ellis, Andy Young and colleagues showed in 1997 that in patients with Capgras delusion their autonomic nervous systems do not respond much more strongly to familiar faces than to unfamiliar faces (much stronger response when a face is familiar is what is normally seen). So a patient expecting to have a strong autonomic response when he sees someone who looks exactly like his wife will not have this experience. What could explain this strange and unexpected (absence of) experience? Well, suppose that this person was not the wife but some complete stranger: that hypothesis would explain the absence of the experience. That’s how James or Maher would explain Capgras delusion using their one-factor theory. But this explanation can’t be correct, because patients with damage to ventromedial regions of frontal lobe also do not respond much more strongly to familiar faces than to unfamiliar faces; yet they are not delusional. They are able to avoid adopting as a belief the hypothesis “This is not my wife”. Capgras patients can’t avoid this, and we infer that this is the case because in the Capgras patients a second factor is present, an impairment of the normal processes of belief evaluation.

The general form of the analysis illustrated above is one we have applied to the explanation of a variety of delusions. It runs like this. First, ask Question 1 (what caused the idea that is the content of the delusion to first come to the deluded person’s mind?). Hope to identify such a potential cause. Then seek evidence from the neuropsychological literature for the existence of patients in whom precisely this cause is present but who are not delusional.  Then claim that what distinguishes the delusional from the nondelusional patients is that the delusional patients have a second impairment, an impairment of the belief evaluation system; this impairment prevents them from using evidence that would prevent them from accepting the belief, i.e. evidence that if used would prevent them from becoming delusional.

Some independent criterion that would support the claim that an impairment of the belief evaluation system characterises all patients with delusions is needed. One approach we have taken here is to point to various pieces of evidence suggesting an association between an impairment of the belief evaluation system and damage to the right dorsolateral prefrontal cortex.

But that’s for another blog post. And so is the topic of hypnotic induction of delusional belief: we work on that with Amanda Barnier, Rochelle Cox and their hypnosis research group.

6 comments:

  1. Two factors: 1: what caused the idea/content to first come to the deluded person’s mind?
    2: what caused this idea to be accepted as a belief?

    Why not three factors? 3: what maintains this belief in the face of contradictory evidence?
    - assumes that for non-delusionals it is possible to no longer believe in something one once did (e.g., santa, weapons of mass destruction, and so on) – i.e., accepting a belief is one act – maintaining it, another.


    ReplyDelete
  2. On behalf of Max Coltheart:

    Yes, this is a good point that needs to be considered. My move to try to avoid this complication is to propose that the cognitive process of belief evaluation that is used when considering whether to accept a candidate belief into one’s belief system is the same process that is used to decide whether new evidence relevant to that belief ought to lead to rejection of that belief. So I would like belief acceptance and belief maintenance to be the same thing i.e. to depend upon the same cognitive process(es).

    There is a very good discussion of exactly this point in a new chapter “Delusion: Cognitive Approaches – Bayesian Inference and Compartmentalization” by Andy Egan and Martin Davies, in The Oxford Handbook of Philosophy and Psychiatry, edited by KWM Fulford et al. A lot of other good stuff in that chapter too.

    ReplyDelete
  3. Thank you for your reply (and the pointer to the Davies & Egan chapter). Your reply helped clarify the propositions – the way that Davies & Egan described the Two-Factor Framework initially glossed over the point I raised, in their telling the questions (you) posed are: 1. What is responsible for the content of the delusional belief? 2. What is responsible for the persistence of the belief? Under this telling question 1 is both about the content of the belief and its adoption; i.e., the second question (about persistence) presumes that the belief has been adopted. In your summary above, you sliced things up slightly differently, phrasing the question 1 in terms of “what caused the idea … to first come to … mind” and the question 2 “what caused this idea … to be accepted as a belief”. Your clarification in the comment made it clear that you are keeping Q1 (about content) and in Q2 lumping together ‘belief acceptance’ and ‘belief maintenance’, i.e., that the process(es) of accepting a belief and maintaining it are the same. This was useful.

    ReplyDelete
  4. Hi Max and thank you for a wonderful post. I'm curious about how you see the relationship between two-factor theory and prediction-error theory. Do you see any areas of overlap between them?

    ReplyDelete
  5. On behalf of Max Coltheart:

    Hello Lisa,

    Thank you for those kind words, and for raising the issue of prediction error, a very important idea here. If by prediction-error theory you have in mind the theory of delusion associated with Paul Fletcher, Chris Frith and Phil Corlett, I think that this account of delusion is wrong because it ignores the kinds of patients we discuss in whom factor 1 is present but there is no delusion: that is, it is a one-factor theory. I also think it is wrong in arguing that delusion is caused by a failure of the prediction-error-signalling system. Our view is that delusions arise only when that system is performing normally.

    Consider Capgras delusion. I go down to the kitchen, expecting, as happens every day, to see my wife there, and when I look at her face I predict (not consciously of course) that my autonomic nervous system will respond. But to day it does not: a prediction error occurs (it only occurs, of course, because my the prediction-error-signalling system is functioning normally). The occurrence of a prediction error is what triggers the search for a new belief, since my current belief system has to be updated (since it characterizes the world incorrectly; it led me to a prediction that turned out to be false). What’s triggered is the use of abductive inference. That comes up with some hypothesis that would explain the absence of the autonomic responses. Acceptance of that hypothesis is acceptance of a delusional belief. But as I explained in my post, there are patients to which all this happens, but who do not become delusional. That’s why we consider that two factors are necessary.

    The same prediction-error story can be told about various other delusions.

    So prediction error is critical to the two-factor theory. For all the different kinds of first factors, the impairment that is the first factor generates a prediction that is not met i.e. the prediction-error-signalling system (which is intact) detects a failure of prediction. That triggers the search for a new belief about the world, a belief that is capable of explaining the event that the patient failed to predict. That belief ought to be rejected, and this will happen if only Factor 1 is present. Factor 2 is needed to explain why the belief is accepted rather than rejected.

    ReplyDelete
  6. Hi all

    There's a reply to this post and comments by Phil Corlett here: http://imperfectcognitions.blogspot.co.uk/2014/03/a-reply-to-max-on-explaining-delusions.html. We posted it separately as it was too long to appear as a comment!

    ReplyDelete

Comments are moderated.