Dear SPMers,

Low Resolution Brain Electromagnetic Tomography (LORETA) is a relatively
new 3D-EEG method that computes the distribution and strength of
intracerebral electric sources from scalp-recorded EEG assuming the
smoothest of all activity distribution (i.e. that neighboring
neurons/voxels show maximal correlation). LORETA computes the spectral
power for every voxel in the brain. Its solution space are 2394 cortical
(!) voxels from the MNI-305 brain. LORETA has been shown to be capable of
correct localization, although with blurring. 

Now, I have an interesting data set (specified below) of co-registered EEG
(from which LORETA can be computed) and [H2O]-PET. This data set is
certainly a valuable basis to compare results of PET and LORETA, and that's
what I want to do.
My idea is to use exactly the same SPM analyses for both domains so that
any differences in the results between the two methods cannot be accounted
for by statistics. This would make the comparison very strong.

My general question to you is therefore:

- do you think the LORETA data could and should be analyzed using SPM? 

Some specific questions are:

- does SPM make assumptions that do not hold for LORETA? 
- does the smaller whole-brain-volume of LORETA (2394 voxels) matter?
- any reasons why LORETA volumes would not conform to random field theory?
- Is the Analyze-Format available so that LORETA images could be
transformed into an SPM-readable format? 

If LORETA can be analyzed in SPM, there are questions regarding the best
way to do the PET-LORETA comparison. My data set comes from the following
experiment: [H2O]-PET and EEG were co-registered in 11 subjects during
placebo or MDMA, both during performance of a visual Continuous Performance
Task (CPT) or a control task (yielding a total of 2x2=4 conditions). There
were 2 PET scans/EEG recordings during each the control task and the CPT.

The first thing that comes to mind is to do an ANOVA over all subjects for
both methods separately and compare the results afterwards (i.e. whether
and where there is an overlap in the patterns of activation and deactivation).
But since both methods are co-registered, i.e. mapping the *same* brain
states, one might argue that it would provide more information to compare
them for each subject separately, without group statistics. What do you think?

Any help you could give me with this analysis is highly appreciated!

Alex

University Hospital of Psychiatry
Research Unit
Lenggstr. 31
8029 Zurich
Switzerland

Email: [log in to unmask]
Phone: +41 1 384 26 32
Fax: +41 1 384 33 96



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%