Guenther Grabner1,2, Siegfried Trattnig1, Markus Barth2
1Department of Radiology, Medical University of Vienna, Vienna, Austria; 2Donders Institute for Brain, Cognition and Behaviour, Radboud University Nijmegen, Nijmegen, Netherlands
A phase model based on 27 individual Parkinson disease (PD) SWI phase data sets was developed which reduces the workload for ROI definition and should reduce intra- and inter-observer variability. This resulting high SNR phase model could be deconvolved reliably using a filtered deconvolution which significantly reduced the influence of the dipolar phase pattern and improved definition of ROIs near tissue boundaries. Phase differences increased by a factor of two in SN and GP in PD patients compared to values in healthy volunteers and were more accurate using the phase model compared to assessment on individual data sets.