Beamforming

Discussions
1

Can I use Beamforming in Brainstorm??

2

Yes, I have updated the source estimation for 2016 to include LCMV beamforming. I have not finished updating the tutorial to explain exactly how to use it. This version uses the “data covariance” calculation in a manner highly analogous to the “dipole modeling,” so you might give both a try.

Beamformers are more exacting about how you calculate the data covariance, as opposed to the noise covariance, so that relatively minor errors in modeling assumptions cause the beamformer to not work, particularly (and ironically) in high SNR situations.

The tutorial I am finishing will demonstrate this effect using the Neuromag Phantom data again, as has been used in the dipole modeling found in “Other analysis scenarios” -> “MEG current phantom (Elekta).”

  • John
3

Thanks a lot sir for the quick reply…
I’ll try to apply and will come to you. It would be great help if you can provide any ‘raw’ tutorial meanwhile so that I can start the test.
(amit.jaiswal2@elekta.com)

4

You can find many example tutorials with different datasets in the section “Other analysis scenarios” of the tutorial page:
http://neuroimage.usc.edu/brainstorm/Tutorials

5

Thanks a lot…
I am going through this…

6

Dear John and Francois,

Since the documentation of BF is not yet ready, it is good to mention that the ‘data covariance’ has to be computed in advance otherwise, BF-LCMV does not appear as an active button in Compute source [2016] menu.

Thanks,
Vahab

7

Thanks, Vahab. I’m updating the Tutorial #22 Source Estimation for all three source methods, then the Advanced Tutorials -> Source modeling -> Dipoles and Beamforming sections.