Due to the generous support of Dennis Molfese, I've been able to make
a significant upgrade to the EP Toolkit, which is intended for ERP
researchers as an adjunct to EEGlab. The main development is that
I've added a graphical user interface to make it easier to use. I've
also added features to further three goals:
1) Facilitating analysis of noisy ERP data, such as developmental or
clinical, by the application of automated artifact correction and
robust statistics at both the averaging stage and the inferential
2) Facilitating multivariate decomposition of ERP data, preferably
using a combination of EEGlab's ICA rotation for the spatial mode and
the Promax PCA rotation for the temporal mode.
3) Facilitating inferential data analysis of ERP data using robust
statistics not previously available with a user interface.
The tutorial provides in-depth guidance on how to use the Toolkit.
This release is currently in beta (meaning no known bugs but it is
still undergoing testing and there are likely further bugs lurking).
If you are interested in trying it out, you can download it from the
This version of the EP Toolkit is more computationally demanding so it
requires Matlab 2006 or more recent. It also needs at least 2GB of RAM
and preferably 4GB or more. In general, the faster the computer, the
better. It has been extensively tested on OS X and is currently
undergoing testing on Windows XP. It can read text, EEGlab .set, EGI
simple binary and EGIS, Biosig EDF, and Neuroscan AVG files. I'd be
happy to implement support for other file formats if a user is willing
to help test them. The EGI format files it saves can in turn be read
by programs like EEGLab and BESA.
In order to use Repov's new OS X compiled runICA code, the attached
binica.m file would need to be used to address a bug that otherwise
can crash the session. Likewise, to use .set files produced by EEGlab
after importing the files from other formats, the following attached
popfileio.m file would need to be used so that the .set files are
correctly formed and thus readable by the EP Toolkit. The changes
have been submitted to Arno and hopefully (if they meet with his
approval) will be incorporated into a future EEGlab release (not yet
incorporated as of eeglab7_1_4_15b).
If you find any bugs, please let me know ASAP so I can fix them. If
making a bug report, make sure to provide full information, including
the error message (if any), the version of the operating system and of
Matlab, and a copy of the data file you were trying to analyze.
Suggestions and general feedback are also welcome. Note, though, that
I don't have grant support for this work so I'm not as free to respond
as folks who have full-time funding to support software development.
Senior Research Scientist
Center for Advanced Study of Language
University of Maryland , Box 25
College Park , MD 20742-0025