ICA - CNL
MATLAB Code and C-code

  • Infomax Algorithm by Tony Bell (Interval Research) and Terry Sejnowski
    Classical Infomax algorithm, It is well suited to separate mixtures of super-Gaussian sources. Most natural sources have super-Gaussian density.

  • Extended Infomax Algorithm by Te-Won Lee, Mark Girolami and Terry Sejnowski
    ICA algorithm to separate mixtures of sub- and super-Gaussian sources. (V2.0)

  • Matlab ICA toolbox for EEG analysis by Scott Makeig, Colin Humphries, Tzyy-Ping Jung, et al. -- over 60 analysis and visualization functions for EEG and ERP data: artifact removal, analysis, and plotting.

  • MEX-file and stand-alone C versions by Sigurd Enghoff of the runica() routine included in the Matlab ICA toolbox for electrophysiological data analysis are now available for download. The compiled versions are 3x to 12x faster than running the original script under Matlab, and are highly memory-efficient for large data sets. DEC, SGI, Sun, and Windows 95/NT versions available. The programs implement versions of the infomax and extended-infomax algorithms and can be applied to any type of data.