Critical thoughts on current rodent models for evaluating potential treatments of alcohol addiction and withdrawal.

Despite years of neurobiological research that have helped to identify potential therapeutic targets, we do not have a reliable pharmacological treatment for alcoholism. There are a range of possible explanations for this failure, including arguments that alcoholism is a spectrum disorder and that different population subtypes may respond to different treatments. This view is supported by categorisations such as early- and late-onset alcoholism, whilst multifactorial genetic factors may also alter responsivity to pharmacological agents. Furthermore, experience of alcohol withdrawal may play a role in future drinking in a way that may distinguish alcoholism from other forms of addiction. Additionally, our neurobiological models, based largely upon results from rodent studies, may not mimic specific aspects of the human condition and may reflect different underlying phenomena and biological processes from the clinical pattern. As a result, potential treatments may be targeting inappropriate aspects of alcohol-related behaviours. Instead, we suggest a more profitable approach is (a) to identify well-defined intermediate behavioural phenotypes in human experimental models that reflect defined aspects of the human clinical disorder and (b) to develop animal models that are homologous with those phenotypes in terms of psychological processes and underlying neurobiological mechanisms. This review describes an array of animal models currently used in the addiction field and what they tell us about alcoholism. We will then examine how established pharmacological agents have been developed using only a limited number of these models, before describing some alternative novel approaches to achieving homology between animal and human experimental measures.