Possibly they are referring to Oakes et al (2007): Integrating VBM
into the General Linear Model with voxelwise anatomical covariates,
Neuroimage 34, 500-508, where the authors show that different grey
matter densities account for a considerable amount of inter-group
variability. Correcting for that is supposed to eliminate false
negatives as well as increase your sensitivity for true activations
due to a smaller error term (probably similar to what you already
know). You can do this pretty easily using FSL. See the manual on
http://www.fmrib.ox.ac.uk/fsl/feat5/detail.html
and search for "Number of additional, voxel-dependent EVs".
As far as I know, it has not become a very widespread standard, but
the rationale seems reasonable. It also depends on your view on VBM,
of course.
Hth,
Cornelius
On Fri, Feb 26, 2010 at 4:30 PM, GS <[log in to unmask]> wrote:
> Hello,
>
> I am curious if anyone here has done a voxel-wise PVC on BOLD fMRI data
> before. I am familiar with PVC in SPECT, PET, and ASL perfusion
> measurements, but have not heard it done with BOLD data. A reviewer recently
> criticised a proposal of ours for not having a protocol for voxel-wise PVC
> of BOLD data, and I am stumped trying to find any reference to methods for
> doing this.
>
> We are not using standard space for analysis but rather the 'functional ROI'
> method of localizing a region in individuals in native space. The issue is
> the anatomical structure where the fROI is localized may be a different size
> between our two populations.
>
> Any thoughts or points in the right direction would be greatly appreciated!
>
>
> My final niggling thought is: is PVC of BOLD even necessary when the
> activation maps are difference images between two different activation
> states (e.g., extrastriate cortex looking at house v objects)? I think this
> is a different creature from resting glucose metabolism images from a PET
> scanner, where less tissue looks like hypometabolism, as in BOLD data we
> wouldn't expect CSF voxels to show significant changes in activity with two
> activation states anyway.
>
--
Dr. med. Cornelius J. Werner
Department of Neurology
RWTH Aachen University
Pauwelsstr. 30
52074 Aachen
Germany
Institute of Neuroscience and Medicine
MR Physics - INM4
Research Centre Juelich
52425 Juelich
Germany
::: Please encrypt confidential data :::
|