Re: [Marsbar-users] Questions about gray volume through clusters

[Matthew B. <mat...@gm...>]

Hi,

>  I obtained the raw data first which I assume has not been adjusted for the
> covariates. When I tried comparing the volumes for the 2 groups using ancova
> and adjusting for the covariates using SPSS there were no significant
> differences? Any comments or suggestions? Maybe I am doing something wrong
> in SPSS.

It's true that marsbar makes no attempt to adjust for the factors in
the SPM model, it only replicated the SPM image scaling for session
effects.  You will need to make sure you have replicated the full SPM
design, including any constant columns, to get something similar in
SPSS.

> If I use the SPM design to get the GM values will it control/adjust for the
> various design parameters applied including the covariates? Or will it just
> adjust for global normalization but not for covariates? If I use this option
> do I have to look for the 'y' or do I still export the summary time course
> in my matlab workspace to look at the volumes?

I would export the summary time course for safety.

> Lastly, are the volumes reported in mL?

The scaling is from the SPM design and the images.  I'm afraid I don't
know what units in VBM scans are, I guess unmodulated would be
probability, and modulated a more complex function of probability...

Best,

Matthew