Hi Donald,
Thanks for your input, I appreciate it.
You are right that the question I want to ask if whether the change in behaviour is related to a change in DTI. I first did a categorical contrast where EV1 below consisted of 1's and -1's for Post and Pre, respectively. However, I did not find any changes (even at the uncorrected level). While this could likely be due to a small sample size (n=7, and I will be adding more subjects as the study progresses), I wanted to see if weighting the main effect with each subject's scores, would be a more sensitive approach. If one subject shows more behavioural change, then his/her contribution would be weighted higher, than those who showed less change.
I'm not sure, but I think the suggestion you brought up might be asking a slightly different question - that is, are there white matter changes that increase in FA, for example, as behavioural changes also increase. What do you think?
In any case, if I were to add the main effect to the model below, i.e. another EV with 1's and -1's, would the model be valid then?
Thanks for your help.
Best,
Joyce
________________________________________
From: FSL - FMRIB's Software Library [[log in to unmask]] On Behalf Of MCLAREN, Donald [[log in to unmask]]
Sent: December 16, 2011 5:21 PM
To: [log in to unmask]
Subject: Re: [FSL] repeated measures behavioural score weighting
It sounds like you want to know if the change in behavior is associated with a change in DTI. If this is the case, then you want to subtract your DTI values and subtract your behavior scores, then do a correlation between the two.
I don't think the current model is valid for two reasons:
(1) You haven't included the main effect of condition, you've included the interaction of behavior and condition.
(2) There is not a consensus on how repeated-measure ANCOVA models should be interpreted.
Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Postdoctoral Research Fellow, GRECC, Bedford VA
Research Fellow, Department of Neurology, Massachusetts General Hospital and
Harvard Medical School
Office: (773) 406-2464
=====================
This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED
HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is
intended only for the use of the individual or entity named above. If the
reader of the e-mail is not the intended recipient or the employee or agent
responsible for delivering it to the intended recipient, you are hereby
notified that you are in possession of confidential and privileged
information. Any unauthorized use, disclosure, copying or the taking of any
action in reliance on the contents of this information is strictly
prohibited and may be unlawful. If you have received this e-mail
unintentionally, please immediately notify the sender via telephone at (773)
406-2464 or email.
On Fri, Dec 16, 2011 at 4:51 PM, Joyce Chen <[log in to unmask]<mailto:[log in to unmask]>> wrote:
Dear FSL,
I have a group of subjects with pre and post dti scans and behavioural scores associated with these. I want to weight each subjects pre and post data with their behavioural scores. I'm not sure if the following design would get at what I'd like to do (and I'd probably have to demean EV1 within the pre and post blocks as well)?
Post>Pre Sub1 Sub2 Sub3
EV1 EV2 EV3 EV4
20 1 0 0
18 0 1 0
17 0 0 1
-10 1 0 0
-12 0 1 0
-15 0 0 1
contrast:
c1: 1 0 0 0 (Post>Pre)
c2: -1 0 0 0 (Pre>Post)
Thanks very much for your advice!
Best,
Joyce
|