By concluding that only “positive illusions” are adaptive misbeliefs, McKay & Dennett (M&D) prematurely discount delusions as biologically adaptive. They do not pursue their argument that delusions may resemble the doxastic shear pin, which breaks down to preserve other “more expensive parts of the system” (sect. 10, para. 1). Therefore, they overlook the possibility that delusions may be tied to survival, and thus, by M&D's own criteria, natural selection.

In remaining alert to nourishment, danger, and reproductive opportunities, humans maintain relatively steady contact with their environment in an ongoing perception-action cycle (Fuster Reference Fuster2006; von Weizsäcker Reference von Weizsäcker1950) and attendant action-outcome learning (Dickinson & Shanks Reference Dickinson, Shanks, Sperber, Premack and Premack1995). The neurobiological changes in early schizophrenia, however, disrupt the “binding” processes of perception, agency, and self (Haggard et al. Reference Haggard, Aschersleben, Gehrke, Prinz, Hommel and Prinz2002), reflected in the patient's experience (Mishara Reference Mishara2007b). How then are delusions adaptive by contributing to the patient's survival?

In making our case, we refer to an analysis of instrumental learning that cleaves the processing of actions into two distinct systems, one goal-directed, the other habitual (Daw et al. Reference Daw, Niv and Dayan2005). The goal-directed system involves learning flexible relationships between actions and outcomes instantiated in the more computationally intensive prefrontal cortices. On the other hand, habits involve more inflexible representations of the relations between environmental stimuli and behavioral actions; when a particular cue is perceived, a specific action is elicited irrespective of the consequences. These two systems compete to control behavior (Hitchcott et al. Reference Hitchcott, Quinn and Taylor2007). Delusions are subserved by the striatal habit system because the computationally intensive goal-directed system is impaired – an argument that we feel has consilience with the authors' view that misbeliefs qualify as biologically adaptive to the extent that they maintain the person's functioning while preventing further damage. (We do not comment on the interesting debate whether delusions are beliefs. For one account, see Mishara, in press a).

M&D argue that brain dysfunction underlying delusions involves “breakage” in the belief evaluation system which is “adventitious, not designed” (sect. 10, para. 4). They advocate a two-factor model: (1) a perceptual insult which engenders odd experiences; (2) a deficit in belief evaluation which enables the entertainment and maintenance of bizarre and unlikely explanations for the experience (Davies & Coltheart Reference Davies, Coltheart, Coltheart and Davies2000). Distinguishing deficit (organic-neural) versus motivational (psychological/psychodynamic-defense) approaches to delusions (McKay et al. Reference McKay, Langdon and Coltheart2007a), M&D conclude that “motivation” plays a psychological but not a biological role in the two-factor model. In contrast, we propose a single impairment in prediction-error–driven (i.e., motivational) learning in three stages: (1) delusional mood; (2) delusion as Aha-Erlebnis; (3) reconsolidation. Our model indicates how delusions may be adaptive as a shear pin function by enabling the patient to remain in vital connection with his/her environment:

1. 1. Prior to delusions, a prodromal delusional mood may last for days, months, or even years (Conrad Reference Conrad1958; Jaspers Reference Jaspers, Hoenig and Hamilton1946/1963). The patient experiences increasingly oppressive tension, a feeling of non-finality or expectation. Conrad (Reference Conrad1958) calls this Trema (stage-fright) as the patient has the feeling that something very important is about to happen. Attention is drawn toward irrelevant stimuli, thoughts, and associative connections which are distressing and unpredictable (Kapur Reference Kapur2003; McGhie & Chapman Reference McGhie and Chapman1961; Uhlhaas & Mishara Reference Uhlhaas and Mishara2007). This reflects an impairment in the brain's predictive learning mechanisms, such that unexpected events, prediction errors, are registered inappropriately (Corlett et al. Reference Corlett, Honey and Fletcher2007).

2. 2. The delusions appear as an Aha-Erlebnis, or “revelation” (Conrad Reference Conrad1958), concerning what had been perplexing during delusional mood. In delusions of reference, harmless or accidental occurrences in the environment are taken as referring to the self. Conrad (Reference Conrad1958) calls this a reflexive turning back on the self in which the universe is experienced as “revolving” around the self as middle-point. The delusions are not primarily a defensive reaction to protect the self, but involve a “reorganization” of the patient's experience to maintain behavioral interaction with the environment despite the underlying disruption to perceptual binding processes (Conrad Reference Conrad1958; Mishara Reference Mishara2010). At the Aha-moment, the “shear pin” breaks, or as Conrad puts it, the patient is unable to shift “reference-frame” to consider the experience from another perspective. The delusion disables flexible, controlled conscious processing from continuing to monitor the mounting distress of the wanton prediction error during delusional mood and thus deters cascading toxicity. At the same time, automatic habitual responses are preserved, possibly even enhanced (Corlett et al. Reference Corlett, Krystal, Taylor and Fletcher2009b).

3. 3. Reconsolidation. Forming the delusion is associated with insight relief which stamps the delusion into memory (Miller Reference Miller2008; Tsuang et al. Reference Tsuang, Faraone, Day and Nicholi1988). Each time delusions are deployed, they are reinforced further, through a process of recall, reactivation, and reconsolidation, which strengthens them, conferring resistance to contradiction rather like the formation of motor-habits with overtraining (Adams & Dickinson Reference Adams, Dickinson, Spear and Miller1981). When subsequent prediction errors occur, they are explicable in terms of the delusion and serve to reinforce it (Corlett et al. Reference Corlett, Krystal, Taylor and Fletcher2009b; Eisenhardt & Menzel Reference Eisenhardt and Menzel2007). Hence the paradoxical observation that challenging subjects delusions can actually strengthen their conviction (Milton et al. Reference Milton, Patwa and Hafner1978). In each rehearsal of the delusion in the present instance, there is a “monotonous” spreading of the delusion to new experience (Binswanger Reference Binswanger1965; Conrad Reference Conrad1958; Mishara, in press b) and, as such, it is both fixed and elastic (Corlett et al. Reference Corlett, Krystal, Taylor and Fletcher2009b). For example, we interviewed a middle-aged schizophrenia patient with the intractable erotomanic delusion that a college acquaintance had fallen in love with her and now controls parts of her life. Whenever she thinks of him, she hears a “car beep” or “trips while walking,” i.e., signals intended to inform her that he knows she is thinking about him.

Neurobiologically, this reconsolidation-based-strengthening shifts control of behavior toward the striatal habit system. However, the ceding of behavior from effortful, conscious control is associated with a “mechanization” of experience. Schizophrenia patients delusionally refer to themselves in inhuman terms, for example as “machine,” “computer,” or “registering apparatus” (Binswanger Reference Binswanger1965; Kraus Reference Kraus1997; Mishara Reference Mishara2007a), as if the delusion reflects its own disabling function of flexible conscious processing. Losing the experience as consistent intentional agent (Wegner Reference Wegner2004), the patients nevertheless continue to respond reflexively to the environmental cues incumbent upon them, necessary for continued survival. As complement to such delusions of alien control, however, the healthy individual has the converse “everyday delusion”: She thinks that it is “I” who moves her own limbs. She calls the movement mine although it has its own momentum, automaticity, and finds its own way. That is, the healthy individual “overlooks” the impersonal-mechanical side of her movements in a “counter-delusion” to the patient who is unable to access the personal contribution (von Weizsäcker Reference von Weizsäcker1956). We are no more free from the necessity of “delusions” in our everyday functioning and its intermittent ceding to automatic processes than is the patient with schizophrenia.

Finally, the authors outline Bayesian mechanisms of rational belief formation. We propose that delusions form via the same Bayesian learning mechanisms but we challenge the strict separation between perception and belief upon which two-factor accounts are predicated (Corlett et al. Reference Corlett, Frith and Fletcher2009a; Fletcher & Frith Reference Fletcher and Frith2009; Hemsley & Garety Reference Hemsley and Garety1986; Uhlhaas & Mishara Reference Uhlhaas and Mishara2007). In our account, delusions also depend on aberrations of perception which occur when neuronal noise induces mismatches between expectancy (Bayesian priors) and experience (sensory inputs/evidence), but in terms of the single factor, prediction error.