Using Illusions to Understand Delusions

which remain firmly held despite contrary environmental evidence, and also the presence of generalized cognitive biases such as the bias against disconfirmatory evidence (Moritz and Woodward, 2006), suggesting a less transparent role of prediction error estimation in schizophrenia. Can the findings of reduced susceptibility to perceptual illusions be viewed as conferring a schizophrenia-related behavioral advantage? If so, is this outperformance of healthy individuals by those with schizophrenia best explained by a predictive failure? The findings of our own laboratory provide empirical evidence that predictive abnormalities are not restricted to passive sensation – in accord with Clark’s suggestion of prediction as a universal neural mechanism – but extend to predictions made on the basis of self-generated movement. Individuals with schizophrenia differ from healthy individuals in their ability to estimate the level of force necessary to match a force applied directly to themselves (Shergill et al., 2003, 2005). Healthy individuals over-estimate the level of necessary force in this context; implying that accurate prediction of the sensory consequences of an action (by means of a prediction error generated via an efference copy of that action being sent to the sensory system) results in sensory attenuation. The reduction of this effect in schizophrenia despite improved performance indexes relative failure of this system. A similar failure to adequately predict the sensory consequences of inner speech (itself a category of self-action) has been influentially postulated to underlie auditory verbal hallucinations (Frith and Done, 1988), and receives growing neurophysiological support from tasks comparing the auditory processing of selfgenerated words and the same sounds played back after recording (Ford et al., 2001; Ford and Mathalon, 2005; Simons et al., 2010). A commentary on