White matter hyperintensities and cognition in Parkinson's disease. (2019)
Type of ContentTheses / Dissertations
Thesis DisciplineMedical Physics
Degree NameMaster of Science
PublisherUniversity of Canterbury
AuthorsHalliday, Sophieshow all
Cognitive decline and dementia are common non-motor impairments associated with Parkinson's disease (PD). White matter hyperintensities (WMHs) are vascular MRI ndings that are common in older individuals and have been shown to be associated with cognitive impairment. This study aims to assess the predictive power of WMH volume and spatial location on cognitive impairments in PD using an automated WMH identi cation algorithm and MRI.
A training data set of 40 subjects with high WMH volume were used to assess the performance of 4 automated WMH detection algorithms against gold-standard manually identi ed WMH maps. The optimal algorithm was then applied to a longitudinal PD cohort (PD=207, HC=51) and cog- nitive impairment and dementia prediction were investigated using Bayesian regression modelling and model comparisons. Speci cally, I investigated the relationship between global and regional WMH volumes, global cognitive ability, and individual cognitive domain scores.
Brain Intensity AbNormality Classi cation Algorithm (BIANCA) was the identi cation algo- rithm that resulted in the most accurate and precise WMH maps in our training set. BIANCA was used with optimised parameters to extract WMH maps for the whole cohort. I found no evidence of increased global WMH burden in PD relative to healthy controls, nor any di erence across cog- nitive ability within PD. Furthermore, I found no signi cant spatial relationship between WMH volume and global cognition, however anterior periventricular WMHs (PVWMHs) associated with attention and executive function cognitive domains.
These results suggest that WMHs are not a clinically-relevant biomarker of cognitive impairment in PD. While there is a weak emerging correlation between increased anterior PVWMH volume and speci c cognitive domain dysfunction, resounding support of regional WMHs prediction of cognitive domain dysfuncytion is lacking.