Dear Andy:
We and I presume others have done this type of analysis all the time.
Unless I'm missing something I don't see why it matters when the masking
occurs. The end result is that SPM will, as you note, only show voxels that
are in common across the group. If you put all subjects into one big
analysis this would take place at the first stage. Using the individual
method it takes place at the second stage. SPM does not, by the way, simply
take the mask images from each analysis and include them; it recalculates a
common mask based on the new analysis.
Also the implicit masking at level one masks your time series so why would
you want to remove it? The notes for spm_spm state that the mask.img file
"...is included for reference, but is not
explicitly used by the results section." The results look at the structure
called xM and all the voxels in XYZ are defined as within mask.
Hope this helps,
Darren
At 12:35 PM 9/10/2001 -0400, you wrote:
>Dear SPM group:
>
>We are considering whether to perform level one analyses of fMRI data (ie,
>basic models and contrast images) in a large multi-subject fixed effects
>model or whether to perform these in batch mode at a subject by subject
>level which would be much more convenient. Because we are interested
>primarily in the level 2 random effects results across subjects, it seems
>to me that the main difference would be how implicit masking is
>handled. If we use a large multi-subject model all contrast images have
>the same mask, but if we run the basic model on each individual we would
>later (at the second level) be combining contrasts based on different masks
>and presumable only analyze voxels that are unmasked in all subjects. Does
>anyone have experience comparing these two methods? And, if so, are there
>any differences in results? Finally, is it possible to turn off implicit
>masking at level 1, retain all voxels in the ????_con images and then mask
>at level 2? I'd appreciate any thoughts on this.
>
>Best wishes,
>Andy Saykin
>
>
>
>
>
>========================================================
>Andrew J. Saykin, PsyD, ABPP
>Associate Professor of Psychiatry & Radiology
>Neuropsychology Program and Brain Imaging Laboratory
>Department of Psychiatry - DHMC
>Dartmouth Medical School
>Lebanon, NH 03756
>
>Tel. (603) 650-5824
>Fax (603) 650-5842
>
>email:[log in to unmask]
>
>http://synapse.hitchcock.org
>
>========================================================
-------------------------------------------------------------------------
Darren R. Gitelman, M.D.
Cognitive Neurology and Alzheimer¹s Disease Center
E-mail: [log in to unmask]
WWW: http://www.brain.northwestern.edus
Voice: (312) 908-9023
Fax: (312) 908-8789
Northwestern Univ., 320 E. Superior St., Searle 11-470, Chicago, IL 60611
-------------------------------------------------------------------------
|