On Tue, May 4, 2010 at 3:03 PM, Mette Lindegaard <[log in to unmask]> wrote:
> We are trying to get an overall understanding of LORETA, VARETA and SPM
> algorithms and statistics. We are writing to you in the hope of getting some
> answers about the software SPM, especially SPM8. We have studied a lot of
> articles about SPM and the SPM manual but we are still a bit confused about
> it. Is the SPM software first doing the source reconstruction and then
> afterwards applying statistical parametric mapping to the reconstruction?
Yes, that is correct.
> you know where we can find information about the algorithm used for source
> reconstruction in the SPM8 software?
The main reference is:
K.J. Friston, L. Harrison, J. Daunizeau, S.J. Kiebel, C. Phillips, N.
Trujillo-Bareto, R.N.A. Henson, G. Flandin, and J. Mattout. Multiple
sparse priors for the M/EEG inverse problem. NeuroImage,
There are several additional papers dealing with some particular
aspects of source reconstruction in SPM that might or might not be
relevant for you.
You can find some more accessible explanations in the slides of SPM
courses from various years accessible from
> In the article "Gender related differences in visual and auditory processing
> of verbal and figural tasks", written by Norbert and Ksenija Jausovec they
> use the reconstruction method COH, which should be similar to the smoothness
> used by LORETA. Is it possible to perform the COH reconstruction in SPM
> without doing any statistic parametric mapping?
Yes, it's definitely possible. If you follow the manual (Imaging
source reconstruction chapter) you'll do just source reconstruction.
Doing SPM on the results is a separate step that is actually not
described in the manual specifically in the context of source