The suggestion that systems occupy something closer to a spectrum than a dichotomy makes

this a potentially powerful way to parse deviance but also very challenging. One example is obsessive-compulsive disorder (OCD) (Graybiel, 2008), where insensitivity to outcome devaluation and slips of action were used to test a hypothesis of dominance by a habitual system (Gillan et al., 2011). Patients with OCD (albeit potentially confounded by the effects of their neuromodulatory therapies) showed no deficit in using rewarding feedback to guide action but instead showed both lack of sensitivity to outcome devaluation and increased frequency in slips of action. A similar conclusion has been derived from observations of the two-step task (Daw et al., 2011) in OCD patients, as they, along with substance abusers and binge eaters, showed a lower dependence on model-based Lapatinib ic50 control (V. Voon, personal communication). this website Furthermore, evidence for abnormalities in components of a goal-directed system in OCD, particularly the caudate nucleus, aligns with a suggestion that key manifestations of this condition reflect on overdominance of a habitual system (Maia et al., 2008). A second example is drug addiction (Belin et al., 2009). One influential proposal is that a protracted exposure to addictive drugs recruits dopamine-dependent striato-nigro-striatal ascending spirals (Haber et al.,

2000 and Joel and Weiner, 2000) from the nucleus accumbens to more dorsal regions of the striatum (Everitt et al., 2008). This results in a shift in control from action-outcome to stimulus-response mechanisms, a putative dominant mode of control in drug seeking and drug relapse. What this entails is that a key mechanism underlying the emergence of compulsive drug seeking, as well as relapse into addictive behaviors, is the subversion of control by a contextually dominant habitual mode.

A final question here relates to the consequence of overdominance of a model-based system. Speculatively, we suggest that it might at least be involved in components of the phenomenology seen in psychotic states, such as paranoia, delusions, heptaminol and hallucinations. The latter can be seen as arising when the sort of processes that are associated with building and evaluating a model become sufficiently detached from external input from the world. We observed that boosting dopamine boosts the impact and control of such model-based influences (Wunderlich et al., 2012b) and perhaps this is at least one pathophysiological step. It is worth noting that in the treatment of Parkinson’s disease, boosting dopamine function often leads to the emergence of psychotic phenomena (Yaryura-Tobias et al., 1970). We have provided an inevitably selective Review of the past, present, and future of model-based and model-free control in humans.