Does abnormal striatal reinforcement learning cause dystonia? A behavioural and functional MRI study (Medical Research Institute Dundee) The Dystonia Society’s seed funding for a new research project in 2015/6 was awarded to Dr Tom Gilbertson of Ninewells Hospital in Dundee. Below he explains his project: The aim of this research study is to demonstrate that patients with dystonia exhibit abnormal behaviour when learning from actions which result in gain or loss (reinforcement learning). This form of learning is explicitly dependant upon the neurotransmitter dopamine and the striatum within the network of nuclei that make up the basal ganglia. We would predict on the basis of previously observed abnormalities in striatal dopamine neurotransmission in patients and animal models of dystonia, that a measurable bias in reinforcement learning is common to all forms of dystonia. Using functional MRI whilst patient are performing a reinforcement learning task, we intend to demonstrate that this can be explained by abnormalities in reward based signalling within the striatum as indexed by the fMRI BOLD signal. We wish to then incorporate this experimental data into a computational model of action selection and reinforcement learning within the basal ganglia. This model will be used to demonstrate that action selection conflict (analogous to dystonic antagonist/agonist co-contraction) is an emergent property of the dystonic reinforcement learning bias. Confirmation of our hypothesis would provide a novel pathophysiological framework with which to develop new treatments for dystonia: rather than a primary movement disorder, dystonic posturing could be considered the secondary, emergent consequence of the reinforcement learning bias and resultant striatal “maladaptive plasticity.” Treatments aimed at reversing this bias may therefore generate a new target for therapeutic intervention in dystonia.