One issue with using SAM for source localization is that signal from correlated sources is suppressed. Although two perfectly correlated signals would be suppressed entirely, real brain signals are not perfectly correlated. Nevertheless, there will be some attenuation of highly coupled sources. One solution to this is to estimate the weights for several sources simultaneously. While this has been implemented in the literature for two sources (the "dual beamformer") the mathematics is not limited to two sources. While there is no fixed limit on the number of sources that can be specified, it is suggested that this number not exceed 10.
Because sam_Nwts operates only for discrete sources, it can only operate on a list of targets, rather than a surface atlas or ROI defined by XBounds, YBounds, and ZBounds.
Similar to sam_wts, sam_Nwts can use a single sphere, multisphere, or realistic/Nolte head model. Additionally, the covariance matrix can be regularized by adding noise to the diagonal using the Mu or PropMu parameters.
Currently, sam_Nwts uses only the Global covariance file, ignoring any markers which may be present in the dataset or parameter file.
sam_Nwts -r <dataset_name> -m <parameter_file_name> [options]
The -r flag designates the dataset name (with or without the .ds suffix), and -m designates the parameter file name.
-v Verbose mode, without this flag sam_wts works silently except for error messages -Z Normalize the SAM weights by the projected noise. -h Show help
Required Parameters (in the .param file or on the command line):
Model: Specify the head model CovBand: Bandpass for covariance file TargetName: Specify use of a list of coordinates on which to calculate the weights
MRIDirectory: Specify where to locate the necessary MRI files PrefixLength: Specify characters used for identifying MRI and headmodel files. ImageBand: Bandpass for weights calculation (if not present, defaults to CovBand).
The output of sam_nwts will consist of a Global.nii file, a nifti image with a single X and Y voxel, with the Z direction encoding the number of targets. The time axis will encode the MEG sensors. In addition, a "legacy" .wts file will be created that can be read by DataEditor.