BrainStorm


GUI Help

Linearly Constrained Min Variance Beamformer (LCMV BF)

Sets the parameters for minimum variance beamformer on the cortical surface. (See m-file help files for sourceimaging and lcmvbf)

The beamformer filtering matrix is:

where Gi is the forward operator for each source location i, and CM is the data covariance. Applying this matrix to the data, generates a timeseries on each source location (filter output). By normalizing the timeseries using their baseline standard deviation, we create a normalized output of the beamformer filter.

The neural activity index is defined as the estimated signal power over the estimated noise power:

where CN is the noise variance, and gi is the ith column of the forward operator. The source power is given by the above equation, if we exclude the noise power term.

CHECK BOXES

DATA TYPE
Check appropriate check boxes (MEG/EEG) to select the type of data.

SIGNAL AND BASELINE REGIONS
Tikhonov Regularization Applies Tikhonov regularization on the data and noise covariance. Tikhonov regularization causes bias on the solution, but reduces the affect of noise. Hint: If not checked, the result is equal to pseudo inverse of the covariance(s).

RADIO BUTTONS

OUTPUT FORMAT
Filter Output Detection maps resulted from the narrowband spatial filter for each time point.
Normalized Output Filter output normalized by the standard deviation of the baseline region for each time point.
Neural Index The detection map obtained by normalizing the estimated spatial source power of the data by the estimated spatial power of the baseline.
Source Power The estimated spatial source power of the data.

See the reference [1] on tutorial for LCMV Beamformer Reconstruction for more information on output formats.

TEXT BOXES

SIGNAL AND BASELINE REGIONS
Signal Specify the start and stop points (in seconds) of the data. Hint:Default values are the Start and Stop values from DataViewer window.
Baseline Specify the start and stop points (in seconds) for the baseline. Hint: These forms are active only if you select Normalized Output or Neural Index as the output format.
Tikhonov Regularization Specify the Regularization parameter lambda as a percentage of the maximum singular value of the covariance matrix.

BUTTONS

OUTPUT FORMAT

Go Run the LCMV beamformer algorithm for the specified parameters. Hint: See Tutorial section LCMV Beamformer Reconstruction for use with the tutorial data.
Quit Quit the window.

 


Window List


photo