Hi all,
I was wondering whether anybody could suggest a way of assessing quality of
dtifit tensor fitting, possibly to detect outliers. I faintly remember that
in an article where C. Pierpaoli was a coauthor, chi-square was proposed as a
measure of the tensor quality fitting. There isn't anything such implemented
in FSL yet, so I wonder if it would be too much work to implemented.
Thanks,
Martin
|