mike
if all male, then it may be worth observing how age of all those male-individuals correlate with cortical thickness......
kk
----- Original Message ----- From: mdkruepke@uwalumni.com To: "Kushal Kapse" kkapse@mail.med.upenn.edu Cc: "freesurfer" freesurfer@nmr.mgh.harvard.edu Sent: Wednesday, June 13, 2012 11:49:29 AM Subject: Re: [Freesurfer] cortical thickness with a continuous variable
kk-
the only issue i can possibly see with this is that all my subjects are all male, and the variable i am going to be looking at in a continues way in this analysis is what we would normally use to differentiate the groups based on cut off scores. That is why i was thinking i would need to just randomly ID the subjects and then use the contrast 0 0 0.5 0.5 to look at the effect of the continuous variable regressing out the group (which should have no real effect if it is randomly applied)
michael
On Wed, Jun 13, 2012 at 10:40 AM, Kushal Kapse < kkapse@mail.med.upenn.edu > wrote:
as per my experience in grp analysis with cortical thickness, age and gender.....i would suggest you to use qdec and select the gender as fixed factor and age/IQ as continuous factor...once you do tht, qdec will define the contrast for you to see "correlation between thickness and age, accounting for gender"; "does avg thickness differs between female or male" AND " does thickness--age corelation differ between male and female"......
if u use qdec, there may not be need to define the contrast for your study unless your trying to use command-line mri_glmfit to do grp analysis....
lemme knw if this helps
kk
----- Original Message ----- From: mdkruepke@uwalumni.com To: "freesurfer" < freesurfer@nmr.mgh.harvard.edu > Sent: Wednesday, June 13, 2012 11:33:27 AM Subject: [Freesurfer] cortical thickness with a continuous variable
Freesurfers-
I am looking to investigate the whether cortical thickness varies in regards to a continues variable, while controlling for other continues variables. I am having a little trouble wrapping my head around this as i am unsure how to compare thickness in regards to a continues variable, how the contrast would look, ect. Looking at this example FSGD ( http://surfer.nmr.mgh.harvard.edu/fswiki/Fsgdf2G1V )
GroupDescriptorFile 1 Title OSGM Class Group1 Class Group2 Variables Age Input subject1 Group1 30 Input subject2 Group2 40 I am thinking i may need to randomly define two groups, and then use a contrast similar to 0 0 0.5 0.5 (contrast 4 on the same page). Which asks:
Null Hypothesis: does mean of group age slope differ from 0? Is there an average affect of age regressing out the effect of group?
Am i correct in thinking that if i use this contrast i would be able to determine whether there is a difference in cortical thickness based on any continuous variable (like age, IQ)?
Michael