Perception and decision making in schizophrenia

Schizophrenia is an illness characterised by auditory hallucinations (hearing voices) and delusions, (unusual fixed false beliefs). These rather perplexing symptoms can be viewed as arising from problems in a) assimilating information from normal sensory processes and b) decision making. How do we normally make sense of the world around us? It is clear that we combine evidence from our senses, e.g vision and hearing, and incorporate that with our previous experience to make a decision about the likely cause of any perception. It is possible to model the perceptual and decision making process by using statistical models based on Bayes Theorem; this allows one to weight different aspects of the perceptual experience and prior expectation to predict the ?normal? probability of making a certain decision. I propose to examine the decision making procedure in normal healthy people and create a statistical model of that process during controlled experiments ? using neuroimaging to see how this model is instantiated in the brain. I will then examine the same decision making procedures in schizophrenia and clarify not only, how their decision making is different; but also the underlying differences in brain function. Furthermore, I will examine this model under the impact of both antipsychotic medication and drugs with the opposite effect on the dopaminergic system; and how it alters when symptoms have been successfully treated.

I propose to examine the cortical basis of perception and decision making in schizophrenia using the Bayesian formulation of predictive models. The key positive psychotic symptoms of schizophrenia, such as hallucinations and delusions, can be viewed as an abnormality in the normal process of developing beliefs about the state of the world, on the basis of incoming perceptual information. Bayesian statistics enable the modelling of this decision making; combining the weight of experience, or prior knowledge, with new perceptual experiences. I will define an appropriate Bayesian model based on psychophysical data from my experiments and examine the underlying neural infrastructure using functional magnetic resonance imaging. The experiments will examine different aspects of perceptual processing and decision making that underpin this mechanism and the role of both symptomatic status, and relevant pharmacological manipulation, through dopaminergic perturbation, upon the parameters in the model. The working hypothesis is that the usual combination of evidence and prior knowledge (in the Bayesian framework) is altered in schizophrenia, either by a) an incorrect distribution or parameterisation of prior knowledge or b) the mechanism that combines this prior with current sensory input. This hypothesis predicts that new evidence (stimuli) or existing evidence (the prior knowledge) are incorrectly assimilated in schizophrenia, leading to faulty perception and decision making.