| Summary: | The motor symptoms in Parkinson’s disease only appear after extensive dopaminergic nigral cell loss, suggesting the presence of redundancy or compensatory mechanisms that serve to delay symptom onset and maintain motor function. Previous studies have demonstrated altered activity in a premotor-parietal-cerebellar circuit, frequently interpreted, but not necessarily established to be compensatory. Unfortunately, it is difficult to differentiate compensatory from direct disease-related if a clear relationship between brain activity and motor performance is not rigorously established. Accordingly, the present thesis investigated fMRI connectivity patterns that predicted motor performance in 12 Parkinson’s patients and 11 healthy controls. Subjects performed a manual tracking task employing a rubber squeeze bulb that incorporated different sinusoidal frequencies and varying amounts of visual guidance. Motor performance was then assessed by first fitting linear dynamical systems models such that the desired tracking performance was the input and the actual tracking performance was the output. A feature of the models (damping ratio) was then used as a metric of performance. The group fMRI connectivity networks were derived by a conditional dependence statistical method, which distinguished between direct and indirect connectivity. Damping ratio from the behaviour models was then predicted by the fMRI connectivity strengths using a sparse linear regression method, and a leave-one-out validation procedure.
In both patients and controls, damping ratio could be accurately predicted with fMRI connectivity patterns. In controls, premotor-cerebellar and cingulate connections were associated with increased damping ratio and enhanced performance. However, in patients the strength of premotor-cerebellar and visuomotor connections were associated with improved motor performance, while connection strengths within the default mode network were associated with worse performance.
Simultaneous modelling of fMRI and behaviour is a powerful tool to assess compensatory changes in Parkinson’s subjects. The current thesis provides strong evidence that altered activity in some parts of the premotor-parietal-cerebellar network is, in fact, compensatory as previously speculated, as greater connectivity within this network contributed to maintenance of performance. Furthermore, activation of this compensatory network impairs the ability of the inferior parietal cortex to normally de-activate as part of the default mode network, possibly making patients susceptible to non-informative, extraneous stimuli. === Medicine, Faculty of === Graduate
|
|---|
