Hi all,
We are looking at analyzing cortical thickness across time in a group of children. Ages are 6-16 and children each have one followup visit, between 1-3 years from first visit. We first ran the cross-sectional pipeline and analyzed change that way before coming across the longitudinal pipeline. The longitudinal pipeline finished without errors and both cross sectional and longitudinal surfaces seem good at first pass.
When comparing the results (change in cortical thickness from first to second timepoint), there are large differences from the cross-sectional to the longitudinal processing in some, but not all of the subjects. The differences in the change of the cortical thickness range from 0.005 all the way up to a difference of 0.04, a difference between processing streams that's as big as the greatest difference in cortical thickness seen within a subject. Some participants switch from a positive to negative difference (i.e. now showing cortical thinning rather than thickening as the cross sectional indicated). Additionally, most of these participants show a greater change with the longitudinal pipeline, which makes us concerned that head size changes are influencing the results.
What would be causing such differences between processing streams? Our conclusions from our data will differ slightly depending on which pipeline we use. Other questions and answers from the Freesurfer mail archive seem to indicate that if the pipeline worked it should be fine. Since it seems to work in our case should we just assume the longitudinal did better, or is there a reason we should be concerned that head growth from pediatric participants might have biased the longitudinal pipeline and caused these differences even though it seems to have run correctly? We weren't expecting the magnitude of differences we saw with the longitudinal pipeline, so we want to make sure before moving forward.
Thanks,
Sally
Sally Stoyell
Clinical Research Coordinator Department of Neurology Massachusetts General Hospital sstoyell@mgh.harvard.edumailto:sstoyell@mgh.harvard.edu M: 607-280-9614