[FieldTrip] 2-dipole beamformer

[Haris Tzagarakis]

Hi There,
I have been trying to implement a '2-dipole' DICS beamformer as in for 
example Schoffelen et al 2008 based on the literature and some postings 
on this list.
This is not to use for coherence work but simply to take into account a 
strong source.
In essence, I have been augmenting every element of my precomputed 
leadfield grid with the leadfield of a selected reference location that 
represents the strong source.
Then, at the beamformer level I get a 6x6 csd matrix for every location 
in the grid and from that, I use the 3x3 diagonal martix that 
corresponds to the 'moving/non-reference' dipole for power estimation.

This all seems to work except that the brain power maps I get show a 
preferential attenuation of the signal in the area of the strong source 
(now other sources are stronger) - and in fact the location selected for 
reference seems to be completely silent. I may be misinterpreting the 
technique here but I wasn't expecting that outcome - I thought that what 
would happen would be that my output map would be similar to the 
original although with lower power levels and that the 'extra' 
contribution of the strong source for every leadfield location would 
find itself in the second 3x3 diagonal (when I plot the power of that 
this seems to indeed be the case *but* the area of reference is again 
attenuated). I think I may have failed to interpret or implement 
something correctly here (most likely both!). Am I doing something wrong 
at the leadfield grid level (does the leadfield matrix of the location 
of reference require special treatment for example?) or should I be 
using the 6x6 csd matrix differently?  - or maybe it could be something 
else? I would be grateful for comments from anyone who has tried this 
before.
Best,
Haris

-- 
Charidimos [Haris] Tzagarakis MD, PhD, MRCPsych
Senior Research Associate
University of Minnesota Dept of Neuroscience

office:
Brain Sciences Center
Minneapolis VA Medical Center
Tel:612-467-1363