Tuesday, 12 September 2017
Can Anosognosia for Hemiplegia be Explained as Motivated Self-Deception?
My name is Andrew Sims. Right now I’m working as a post-doctoral researcher at the Université Catholique de Louvain, in an Action de Recherche Concertée (ARC) project on action theory and neuroscience. Before this, though, I wrote my PhD dissertation in the philosophy department at Deakin University and while visiting at the Department of Clinical, Health and Educational Psychology at University College London. My dissertation focused upon the psychodynamic explanation of anosognosia for hemiplegia (AHP)—denial of paralysis—and its relationship to cognitive deficit models of the condition. I’ve recently had an article published in the Review of Philosophy and Psychology that draws upon some of this work (Sims 2017). Thanks very much to Andrea Polonioli for inviting me to introduce it on the Imperfect Cognitions blog.
AHP occurs after brain damage and involves a paralysis on one side along with various attitudinal distortions toward that paralysis. At a minimum, this is an inability to recognize the deficit which ranges from a mild forgetfulness to a consistent delusion that the body is fully functional. However, there are various other attitudinal distortions that also occur: anosodiaphoria, an inappropriate lack of negative feeling about the deficit (though not a flattening of affect in general, as is sometimes claimed, since this is often coupled with a disproportionately negative attitude towards minor unrelated complaints); misoplegia, an exaggerated disgust or hatred towards the affected limb which is felt to be alien and intrusive; and somatoparaphrenia, the attribution of ownership of the paralysed part of the body to another person (like a doctor or relative). Furthermore, these patients sometimes appear to have implicit knowledge of the deficit in parallel with the explicit unawareness.
These attitudinal distortions led earlier clinicians to explain the condition in terms of psychological defense against negative emotion associated with the deficit (Weinstein & Kahn 1950). On this view the attitudes are caused by conative factors (like a desire-like representation of bodily integrity) which distort belief formation. Since then, in part because of an influential critique by Bisiach and Geminiani (1991), these accounts have largely been replaced by deficit theories which propose that the delusion is caused by a deficit in the ability to compare predicted motor outcomes to sensory feedback, possibly also coupled with a deficit to executive function or working memory (e.g., Davies et al. 2005). Prime amongst the objections of Bisiach and Geminiani are two: first, that the defense theories can’t explain anatomical features of the condition; secondly, that they can’t explain why the paralysis is selectively denied (where other deficits—like cortical blindness in one case—are not).
In my paper I argue that these objections are only compelling on implausibly strong versions of the defense theory, and that we should want to integrate features of the psychodynamic account into the deficit theories in order to account for the diverse attitudinal symptoms. The way I argue this is partly by responding to the objections directly and partly by showing what a hybrid theory of AHP might look like. With respect to the former, I show that defense theories need not exclude neurological factors from playing a causal role in the condition, and furthermore, that including a neurological factor will help account for the selectivity of the anosognosia. Dual or multiple-factor theories of delusion are drawn upon in the context of deficit theories in order to explain why a delusional belief of a specific kind is formed, and there is no reason why this approach is invalid in the context of defense theories (Aimola Davies et al. 2009).
With respect to the latter, I discuss the work of Fotopoulou (2013) and Turnbull et al. (2014), who make sense of deficit accounts for anosognosia within a predictive coding framework and in terms of emotion regulation. In this framework, motor commands are construed as prior beliefs in a hierarchical generative model which drive action by causing the selective sampling of the appropriate sensory (e.g., proprioceptive) states. Fotopoulou suggests that AHP can be understood under a predictive coding framework in terms of prior beliefs about the state of the body being given an aberrantly strong weighting, such that they aren’t updated when they encounter prediction error. Turnbull et al. (2014) also suggest that this recalcitrance of prior beliefs is associated with the regression to a (both spatially and emotionally) egocentric view of the world, such that the patient has trouble becoming aware of facts that are emotionally unpleasant.
There are two features of their mechanism sketches which are particularly amenable to defense theories. The first is that—given the interpretation in this framework of conative states as prior beliefs—there is no great problem in understanding how conative states could influence belief formation. That is because the difference between doxastic and conative representations is one of degree rather than kind: a representation is conative to the extent that it drives action rather than perception, but this can change depending on its overall role in the economy of prediction-error minimisation. So conative states can interact with belief formation processes in just the same way that doxastic states can. The second is that it allows for some individual variability in the way that the deficit is denied (e.g., anosodiaphoric vs. misoplegic strategies) since it leaves a role for individual variability in the premorbid prior beliefs which become recalcitrant.
The move away from explanations in terms of motivated defense is premature, given the power of this style of explanation for the attitudinal distortions associated with AHP. It is just that we need to be serious about understanding how these processes might work at the level of mechanism, and work out how the associated hypotheses can be made empirically tractable.