Apathy and impulsivity are common and debilitating consequences of many neuropsychiatric diseases, including Parkinson's disease. Despite their prevalence and prognostic importance, their underlying mechanisms remain poorly understood. This knowledge gap presents a barrier to clinical assessment and treatment. Previous studies have focused mainly on dopaminergic influences on apathy and impulsivity, suggesting dopaminergic opponency. However, noradrenergic systems are also associated with the cognitive and motivational processes related to apathy and impulsivity.

In this thesis I begin by outlining a new approach to the study of apathy, using Bayesian models of brain function. These models cast behaviour as an inferential process, where sensory information is continuously integrated with predictions or 'prior beliefs' about the consequences of action. I propose that apathy is a failure of active inference, caused by aberrant precision of prior beliefs. I hypothesised that apathy is associated with reduced precision of priors for action, and that the relative weighting of priors is modulated by noradrenaline. To test these hypotheses, I developed a visuomotor task that requires effortful, goal-directed behaviour, and used Bayesian modelling to infer the relative precision of priors.

I first applied this approach in a healthy young cohort, focusing on individual differences in trait apathy. Apathy was negatively associated with model estimates of prior precision, suggesting that more apathetic individuals had less precise priors for action. I then tested the generalisation of these results to (sub)clinical apathy in elderly participants and people with Parkinson's disease.

The people Parkinson's disease additionally underwent a pharmacological study with the noradrenergic reuptake inhibitor atomoxetine. The baseline integrity of the noradrenergic system was quantified by ultrahigh field 7T MR imaging of the locus coeruleus. I found that the effect of atomoxetine on patients' prior beliefs varied strongly as a function of locus coeruleus integrity. Specifically, the patients with a more degenerate locus coeruleus benefitted more from treatment by atomoxetine.

In the psychopharmacological study, I also measured impulsivity in participants, in terms of their stop signal reaction time (SSRT). Compared to elderly controls, Parkinson's patients had a longer SSRT (more impulsive), with no differences in task accuracy or choice reaction time. Crucially, the effect of atomoxetine on SSRT depended on locus coeruleus integrity, such that Parkinson's patients with a more degenerate locus coeruleus showed a greater improvement in SSRT on Atomoxetine.

I conclude that apathy can be understood as a failure to assign the necessary precision to priors for action, with that precision moderated by the noradrenergic system. This Bayesian account paves the way to a common understanding of apathy across clinical disorders, to examine its relationship to impulsivity through common noradrenergic impairments, and to serve as a target for novel treatments.