Abstract

Obesity has often been thought of primarily as a metabolic condition but there is an increasing tendency to consider it as a “neurobehavioural syndrome with metabolic consequences”. This has led to a corresponding growth in the application of cognitive neuroscience approaches, particularly functional neuroimaging, to understanding over-eating and weight gain. However, the results of such studies have proven inconsistent and accompanying brain-based models of obesity, such as food addiction, or altered reward processing, have, when the literature is viewed as a whole, received little convincing support. To me, the overwhelming message of this growing literature, is that there are many, interacting causes of obesity and the challenge for the field is to begin to comprehend this phenotypic heterogeneity. Current ideas of brain function have come to emphasise its participation in the overall effort of maintaining the stability of the organism within an environment that is both generally predictable but also, at differing timescales, complex and changing. By learning about, predicting and capitalising on environmental regularity, the system as a whole survives and thrives. This perspective – essentially one of the brain as a homeostatic organ- offers a powerful framework for thinking about how it contributes to appetite, eating and weight control. We can conceive the complex array of appetitive behaviours as emerging from the brain’s integration of an array of interoceptive and environmental signals and, as such, a comprehensive understanding of the brain’s role in obesity cannot be developed without reference to these signals, to how the brain integrates them and to how it uses them to select and motivate highly complex (and, sometimes apparently irrational and harmful) behaviours. I will aim to show how we are beginning to link traditional homeostatic views of eating with computational and cognitive concepts drawn from neuroscientific studies of learning and decision-making. I will illustrate this by presenting data from behavioural, psychopharmacological and neuroimaging studies.