Latest version 

Java-based Exploratory Data Analysis for EEG

This Java-based application can be used for performing a range of signal processing operations on EEG data files, up to and including visualization and scoring of spectra and ERPs.

It can be used for analysis of EEG recordings obtained from several sources, including files in EDF format. This program has a special relevance to members of BRAINnet — the group of researchers with a common focus on the wide-ranging body of brain function data provided through BRAINnet. Jeda is the means to perform fully-customized analyses on EEG files in the BRAINnet database.

The problem with EEG analysis is not that it is difficult, but that it is awkward. It is awkward due to the multitude of formats and the diversity of analysis strategies for time-series data, both of which discourage the development of standard analysis tools. The present application addresses these problems by providing a framework for EEG analysis, which hard-codes routine jobs while relying on scripts and plug-ins to accomodate innovation.

The scripts and plug-ins help to modularize the problem of EEG analysis; and help in testing, comparing and sharing of methods.

The application is not for everyone. Its size and wide-ranging goals result in a complexity that may be offputting. However its modularity and batching capability should appeal to anyone involved in specialized and/or large-scale EEG processing.

There are presently two versions of Jeda available:

1. The 'detailed' version, which requires compilation and comes with only a few datasets: see Introduction to Jeda
2. The 'demonstration' version, which is pre-compiled and comes with a large number of datasets, but is minus certain complications: see below

The most up-to-date version of the documentation can be found at www.brain-dynamics.net/jeda/index.html

Documentation

Demonstration of Jeda

An easy way to see Jeda in action is to:

1. Confirm that you have Java installed, that the version is 1.5.0 or later, and that certain environment variables are set. See Java installation instructions for details of how to do this.
2. If you have the demonstration version on a CDROM, then Window's 'autorun' feature means that simply inserting the disk should cause the program to run.
3. If you have copied the demonstration version of Jeda to your hard disk (or 'autorun' fails, or you have some other version of Jeda on hard disk), then you will need to run the program manually:
   • Open a terminal window
   • Change to the directory containing the distribution
   • Run the demonstration by typing ./demo (if Linux) or demo (if Windows). (This runs a batch file, demo or demo.bat, which in turn runs Java.)
4. A small window 'Demo of Jeda' will appear, which provides a simple GUI front-end to Jeda. Choices can be made, and pressing 'Proceed' will cause various graphics various graphics to be generated.

Note: Launching the demonstration program manually via a GUI-based file browser is also possible. However this method precludes the use of extra command line switches, and prevents easy viewing of error messages. If manual running of the program is necessary (and it is necessary for the detailed version), then some familiarity with command line operations is advantageous.

The GUI allows choice of

• Either of two subjects: 81237443-1 or 81567509-1
• One of three referencing options: A1A2, Av or Cz
• Paradigm: EO,EC,AH,ODDB,GNG,EYET,LET,WM,MAZE,PPI, or FACES
• View: Summary, Detailed, Spectra or ERPs

A full description of the eleven paradigms mentioned above is available to BRAINnet members. Visit BRAINnet to learn about how to become a member.

You must have the 22 raw data files (417MB) installed for this demonstration to work. Look for a subfolder 'data' containing files named '81237443-1.*.NS5' and '81567509-1.*.NS5'. The files are part of the demonstration version. If you have some other version of Jeda, the 22 data file can be obtained as a separate download.