[Eeglablist] multiple components from the same subject in a cluster

[Fran Copelli]

Dear list,

I'm looking for either steps on how you approach this problem, or
instructions on how to use the version of kmeans that forces one component
per cluster (see text from EEGLAB wiki below).

"In general, when multiple components from the same subjects in ICA
clusters becomes a problem, we prefer either (1) to use at most 1 component
per subject per cluster because this avoids having to compromise with the
statistics (this is possible when using the CORRMAP plugin for clustering
data; *there also exist a** version of kmean that forces to use one
component per cluster)* or (2) remove components manually in clusters."

Thanks very much,

Fran

SMART Lab, Psychology Department