Dear Professor Stephen Smith

Thanks for your constructive response. I have some further questiones, is it valid if i perform halfway space registration for some subjects with 2-point scans and some subjects with 3-point scans in the same study? one thing i would concern is about the degree of smooth of the half-way space template, it seems the smoothness of the individual half-way space template is different across subjects. In addition, i have some experiences on customizing the image preprocess scripts using FSL command, could you provide me some cues (or demo scripts) for halfway registration for TBSS procedure (currently i have no idea about which command and pre-process procedure should i use in such kind of analysis)?

Many thanks and hope hear the feedback from you and all FSL experts 

Best

Paul

Date: Sat, 26 Jan 2013 07:41:57 +0000
From: [log in to unmask]
Subject: Re: [FSL] The optimal way for longitudinal TBSS procedure with FSLV5.0.2
To: [log in to unmask]

Hi - there's a chance that B could be more sensitive, but currently TBSS isn't setup to make that a trivial option, so you would need to do a lot of script customisation yourself.  The easy and safe way forwards is A.

Cheers.


On 25 Jan 2013, at 02:44, Paul Chou wrote:

Dear all FSL experts

Currently, i want to analysis the 3-yr longitudinal DTI dataset with TBSS protocol. After searching the PUBMED and FSL archive, i found there are two major approach for such kind of analysis: A. treated the longitudinal data as cross-section data and perform longitudinal based statistical permutation analysis; B. using half-way space (using FLIRT and midtrans?) registration scheme for the longitudinal DTI data and also perform longitudinal based statistical permutation analysis. Because of some clinical issues, it is hard to let all the subjects have 3 time-point scans, in the current study design, i just include the subject at least have 2 time-point scans into the analysis procedure (of course, most of subjects could have 3 time point scans !). Based on above situation which i faced, which analysis pipeline would be most optimal for this study? Any helps and related opinions were welcome !

Many thanks

Best

Paul



---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director,  Oxford University FMRIB Centre

FMRIB, JR Hospital, Headington, Oxford  OX3 9DU, UK
+44 (0) 1865 222726  (fax 222717)
[log in to unmask]    http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------