Jenna Gelfand wrote:
> Dear SPMers,
> I have functionally defined a region of interest mask using SPM, and I now wish to
> extract the mean signal intensity at each time point within the time series, averaging
> across the voxels within the ROI. After converting mean signal intensity values to
> percent change scores for several subjects, the results do not fit with my hypothesis
> or with the SPM RFX results using the GLM. I am using in-house tools to extract the
> mean signal intensity from each subject's pre-processed data. I'm concerned that
> lack of global scaling is the problem. I tried to normalize all signal intensity values to
> the grand mean of all signal intensity values across all subjects,
The default scaling in SPM is
within-session - see below ...
> but that seems to be
> a rather crude method and I am skeptical of the results it produces.
>
> Any suggestions? I would be so appreciative!
>
The default scaling in SPM is as follows
(i) Compute the within session global
value, g, as the average value at all
voxels at all time points in that
session. (ii) Then define a 'brain
voxel' operationally as one
having a value more than g/8. (iii)
Now recompute g just based on brain
voxels, (iv) multiply all voxel values
at all time points by 100/g.
Am not sure if this helps ??
Best wishes,
Will.
> Regards,
> Jenna Gelfand
> UCLA Brain Mapping Center
>
>
>
--
William D. Penny
Wellcome Department of Imaging Neuroscience
University College London
12 Queen Square
London WC1N 3BG
Tel: 020 7833 7478
FAX: 020 7813 1420
Email: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
|