Hi,
On 27 Feb 2009, at 12:52, Eva Kenny wrote:
> Hi,
>
> I was hoping for some guidance on how I may interpret my Melodic
> results.
>
> Group Analysis:
> I have 2 groups of subjects and entered all subjects into Melodic
> and used
> the temporal concatenation approach (as I have resting-state fMRI
> data). Is
> there a way of comparing the two groups, rather than just by visual
> inspection?
There are various option depending on how you want to define what you
are actually comparing. You could look at melodic_mix, which gives you
one column per ICA component, and in each column is all subjects'
concatenated timeseries associated with that particular component. You
could cut that up into chunks, estimate the temporal variance for each
subject, and do a 2-group t-test on those variances; this will give
you one 2-group test for each ICA component.
Christian has recently been working on a more sophisticated approach
that lets you compare both strengths and voxelwise differences; we
will explain how to do this shortly I think.
> Single-subject Analysis:
> I have also run Melodic single subject analysis and then using the
> template
> matching procedure approach I identified 1 component for each
> subject that
> most closely matched the RSN of interest.
>
> I have tried a few methods to see if I can identify differences
> between
> groups, but I am unsure of the output I get, e.g. randomise- is it the
> ..._max_tstatx file I should be looking at for differences between
> groups?
Depends what you want - if you are looking for p-values corrected for
multiple comparisons across voxels then yes you want "max" either
voxelwise, clusterwise, or using TFCE (see randomise manual).
Cheers.
> And is there an image I can overlay this on. There are very few
> clusters on
> my max_tstat image, however on the tstat and vox _tstat there is a lot
> activation.
> I also tried running the tensor ICA option in Melodic on the
> selected ICs
> but I got error messages - nifti image read.
>
> Any help would be greatly appreciated.
>
> Many thanks,
> Eva
>
---------------------------------------------------------------------------
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
---------------------------------------------------------------------------
|