Dr Andrew Higginson, Exeter : "Learning to be ill? Adaptive decision-making systems with limited information may cause ill health"

Speaker:

Dr Andrew Higginson, Senior Lecturer, Psychology Department, University of Exeter, UK

Title: "Learning to be ill? Adaptive decision-making systems with limited information may cause ill health"

Abstract:

Non-communicable illnesses such as obesity, depression, and anxiety disorder are the dominant medical problems in the developed world, but their causes are poorly understood. Evolutionary explanations are limited to reasons why the deleterious condition could be adaptive and often fail to capture critical phenomena. For instance, explanations for depression focus on depressive behaviour as a way to avoid costly effort where benefits are small and/or unlikely, but fail to explain why low mood persists when the situation improves. Animals, including humans, do not always make optimal decisions but have evolved decision-making systems (or ‘rules’) that are adaptive generally. I will describe an emerging paradigm for understanding illness based on computational models of adaptive learning in which individuals face repeated choices among options that affect fitness and alter beliefs. Among populations of individuals all following adaptive rules a significant minority show behaviour associated with illness. For example, our model of depression predicts 5% of individuals will  be inactive when it would be better to interact with the world, mimicking an effect of depression on behaviour. These explanations for depression, anxiety disorder and obesity contract strongly with the idea that the system is somehow malfunctioning; brains of depressive patients may be working perfectly but have inaccurate beliefs about current conditions. Our work suggests that medical interventions for these illnesses are unlikely to work because they do not adjust beliefs: effective treatments are likely to involve altering the experiences of adaptive decision makers.