In this paper, we consider the concept of food addiction from a clinical and neuroscientific perspective. Food addiction has an established and growing currency in the context of models of overeating and obesity, and its acceptance shapes debate and research. However, we argue that the evidence for its existence in humans is actually rather limited and, in addition, there are fundamental theoretical difficulties that require consideration.
We therefore review food addiction as a phenotypic description, one that is based on overlap between certain eating behaviours and substance dependence. To begin, we consider limitations in the general application of this concept to obesity. We share the widely held view that such a broad perspective is not sustainable and consider a more focused view: that it underlies particular eating patterns, notably binge eating. However, even with this more specific focus, there are still problems. Validation of food addiction at the neurobiological level is absolutely critical, but there are inconsistencies in the evidence from humans suggesting that caution should be exercised in accepting food addiction as a valid concept. We argue the current evidence is preliminary and suggest directions for future work that may provide more useful tests of the concept.