What determines what you will believe? At any instant we are being confronted by a complex environment in which multiple events and objects engage attention and influence choices, beliefs and behaviours. Neuroscientists at Cambridge are currently exploring the psychological and brain bases for motivation and the formation of beliefs in order to understand disorders such as delusional thoughts and binge eating.

Psychologists in the Department of Experimental Psychology at Cambridge and elsewhere have developed laws of learning, which predict the rate and conditions at which animals learn how to obtain food rewards – critical for survival in the wild. Cambridge psychologists showed that these same laws of learning apply also to humans when we learn about causes. Researchers from the Department of Physiology, Development & Neuroscience went on to discover that, remarkably, certain individual brain cells called dopamine neurons also obey these same laws of learning. Scientists have developed a number of concepts that are particularly important in this research, namely ‘prediction error’ and ‘motivational salience’. Prediction error alludes to mismatches that occur when there are differences between what is expected and what actually happens. It is vital for learning. The scientific theory of prediction error learning is encapsulated in the everyday phrase “you learn by your mistakes”. When there is a mismatch between expectations and outcomes, this mismatch (or mistake!) is noticed, and, in order to learn effectively, expectations for the future are revised accordingly. Motivational salience refers to the extent to which a given object, event or thought captures attention and drives goal-directed behaviour. Prediction error and motivational salience are likely to be highly co-dependent and there is strong evidence that both are reliant upon the neurotransmitter dopamine acting deep within the brain in a region called the mesolimbic system.

Neuroscientists from the Department of Psychiatry, in collaboration with the Behavioural & Clinical Neuroscience Institute, recently showed that it is possible to safely study learning processes and prediction error in the human brain using neuroimaging techniques. This work began with the intention of understanding more deeply the delusional beliefs that can occur in some mental illness. The starting premise, that a key factor in the formation of these irrational beliefs lies in aberrant learning of associations, was borne out by observations that the behaviour of the system is predictive of whether a person will experience delusions both when exposed to a pharmacological challenge or when mentally ill. Patients with schizophrenia show clear disruptions in the sensitivity of the frontal and mesolimbic areas of the brain to prediction error and this correlates with the person’s propensity to delusions. Recent work on psychotic illness has linked deficits in the mesolimbic system to changes in motivational salience, suggesting that people who suffer from paranoid ideas do indeed over-attribute importance to chance events and coincidences.

Thus core neuroscientific concepts of prediction error and motivational salience in the mesolimbic system are useful tools in developing our understanding of how adaptive, and maladaptive, beliefs are formed. However, it appears that the implications may be even more farreaching. In parallel studies on healthy and obese individuals, it is clear that the same brain system is highly sensitive to motivationally salient items, such as pictures of food. Importantly, the sensitivity of these systems is modulated by the individuals’ background state – a hungry individual shows a higher degree of mesolimbic sensitivity than a sated one. Moreover, it appears that a deficiency in leptin, a key appetitemodulating chemical, disrupts responses in this system. The emerging pattern is an interesting one: it appears that basic motivational and learning processes could, when malfunctioning, be at the heart of a range of abnormal behaviours, from delusions to harmful food intake.