Hi
> First of all I want to compare resting state data of two groups of
> subjects…I thought I had to do the group analysis but if I’m right
> then you
> get only “mean” components for the whole group and no individual
> components
> per subject.
That's right
> So do I have to do single subject analyses and compare the
> components of interest from these analyses?
>
If you have two groups then you can use group analysis twice and
compare the outputs qualitatively. We're currently finishing off an
alternative way of doing this quantitatively but for the moment this
seems the best option - or run individual ICAs and compare e.g. using
the approach described in Mike Greicius' 2004 PNAS paper.
> Second to select the right component I already saw some earlier
> discussions.
>
> Christian answered to do it this way:
> you can calculate spatial correlations between maps using the
> command line
> tool avwcc. I think the analysis in the paper was done within
> matlab, you
> can quite easily replicate it, again using the avw tools, e.g. if
> you have
> a binary template use
>
> avwmaths template -mul ICmap ICmap_inside
>
> and
>
> avwmaths template -mul -1 -add 1 -mul ICmap ICmap_outside
>
> and then use avwstats to look at the mean score (use -M to get the
> mean over
> non-zero values only)
>
> it was already some time ago and I saw avwcc is now fslcc, so I
> don’t really
> know how to do it.
only the name has changed, so use as before, simply call fslcc between
two melodic_IC sets (make sure the maps are in the same space!)
>
> Then finally to compare the two groups; In Matlab there is the
> possibility
> to do a full factorial design to compare different components
> between the
> two groups is their a similar option in fsl or do I need to do it in
> Matlab?
easiest is to use randomise on the full set of maps (after collecting
these into a single 4D file using fslmerge -t). You can then test for
within or between group effects using a design matrix (which you can
create using Glm - or GLM_gui on the mac) and a contrast matrix (dito)
> And is this possible to use fsl output in Matlab for second level
> analysis?
Yes, that too - use the scripts in $FSLDIR/etc/matlab to read these
maps in
hth
Christian
>
> Thank you very much,
>
> Elisabeth Jonckers
_______________________________________________
Christian F. Beckmann, DPhil
Senior Lecturer, Clinical Neuroscience Department
Division of Neuroscience and Mental Health
Imperial College London, Hammersmith Campus
Rm 419, Burlington Danes Bldg, Du Cane Road, London W12 0NN, UK
Tel.: +44 (0)20 7594 6685 --- Fax: +44 (0)20 7594 6548
Email: [log in to unmask]
http://www.imperial.ac.uk/medicine/people/c.beckmann/
Senior Research Fellow, FMRIB Centre
University of Oxford
JR Hospital - Oxford OX3 9DU
Tel.: +44 (0)1865 222551 --- Fax: +44 (0)1865 222717
Email: [log in to unmask]
http://www.fmrib.ox.ac.uk/~beckmann
|