Dear Mark,
This was the earlier comment and the relevant sentence (https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1206&L=fsl&P=R36398&1=fsl&9=A&I=-3&J=on&X=DEFFAEAC5E9F233271&Y=perlaki.gabor%40freemail.hu&d=No+Match%3BMatch%3BMatches&z=4) "......... However, if this is not sufficiently good but your non-linear registrations are good then it is possible to extract an approximate linear registration from the non-linear result, which would then allow you to use the FIRST method without my above concerns about statistical bias. ......"
Thanks,
Gabor
|