there is no problem in using DCM with 32 channels. This is my personal
experience; no one has done a formal comparison of using DCM with
different number of channels or montages.
Keep in mind that DCM can be seen as a source reconstruction exploiting
temporal constraints. That means, with respect to the spatial resolution
of sources, you face the same issues as conventional source
reconstruction of a single time point. In particular, typical setups for
EEG (sensors on the upper half of the head) are probably suboptimal in
resolving source z-direction parameters in the temporal lobes (both
orientation and location). In other words, it depends mostly on the
experimental question you want to ask, and whether your sampling allows
to resolve the crucial parameters (both time and space). I personally
would, if in doubt, use as many channels as possible.
All the best, Stefan
> Is there an optimal number of EEG channels for a DCM analysis?
> We're thinking of using either 32 or 64 channels for recording. It would, of
> course, be time-saving to use only 32, but we're worried about not getting
> the best results with 32 channels.
> I'd be thankful for your suggestions!
> Johannes Klackl, University of Salzburg
Dr. Stefan Kiebel
Wellcome Trust Centre for Neuroimaging
Institute of Neurology, UCL
12 Queen Square
London WC1N 3BG
Phone: (+44) 20 7833 7478
Fax: (+44) 20 7813 1420