 |
 |
The limiting factor is probably the registration between functional and structural images. Skull stripping and bias correction of anatomical images sometimes helps, but the main thing to sort out is the nonlinear distortion in the EPI data. The field mapping toolbox can be of help for this, but it requires additional data to be collected. Another possibility, which works for some data, is to do the segmentation and Dartel registration directly on the fMRI. Carlton, my former PhD student, uses this approach quite a lot, and finds that it can work fairly well. However, if your fMRI data have very large dropout artifacts, the results may not be quite so good. Best regards, -John On 20 April 2011 16:31, Simon Vandekar <[log in to unmask]> wrote: > Hi DARTELers, > > I am using DARTEL to normalize functional images of adults between 60 and 85 to the MNI template. attached is an example of the result. > I am finding that there are areas where functional images do not normalize well to the MNI, in this example the visual/cerebellum area, but in some cases in the PFC and around the ventricles. > > 3 questions: > > Does anyone have a suggestion of how I might improve the quality of normalization by changing parameters of DARTEL? > Will skull stripping before aligning all the functionals and structurals yield a worthwhile improvement? > Is there anything I can do to improve normalisation or is it a limitation of trying to normalise this age range to the mni? >