EEGLAB exists as a compiled binary for Mac, Windows, and Ubuntu and may be downloaded on the EEGLAB download page. The video below shows how to install the compiled version of EEGLAB.
Table of contents
- Starting EEGLAB
- Similarity between the compiled and the MATLAB version of EEGLAB
- Frequently asked questions
- How to check the integrity of the compiled version
- Can I compile EEGLAB myself?
- Download the ZIP file for the EEGLAB compiled version on the download page and uncompress it
- Each zip file contains 2 folders “for_redistribution” and “for_redistribution_files_only.” Start by starting the installer in “for_redistribution” folder. If all goes well, this is the only thing you have to do to install EEGLAB.
- If for some reason, the installation process fails (permission issue, for example), you may use the files in the folder “for_redistribution_files_only.” Note that these files require that the MATLAB runtime engine be installed on your system - this step should have been done automatically above but we will assume here that this step fails. These are available as separate download here (the Runtime version you need should appear when you attempt to start EEGLAB).
- Notes for MacOS and Linux users: Apparently, the compiled version of EEGLAB needs to be installed using root user credentials. If you experience problems during the installation process on Mac, after downloading the zip file, open a terminal, go to the folder for_redistribution/MyAppInstaller_web.app/Contents/MacOS and type in sudo ./setup. This will start the installation as root. Once installed, you may use EEGLAB under your user (non-root) account.
- Click “EEGLAB” to start EEGLAB (however, see below)
- Create shortcuts if necessary on the Desktop and in the “Start” menu
- It is often preferable to have the command line output of EEGLAB as it contains relevant information. On Windows, start EEGLAB using the file “eeglab_run_this_one_on_windows.bat.” On OSx and Ubuntu, use the file run_EEGLAB.sh (might require to have its permission changed to be executable; otherwise, the content of the script will be shown).
- Both version graphical interface are identical
- There is nothing in the EEGLAB graphic interface that you can do under MATLAB that is not possible to do in the compiled version of EEGLAB. This includes using all the plugins and all the external modules attached to EEGLAB, saving scripts, and running MATLAB scripts.
- It is not possible to add new plugins or download and install third-party plugins. To do so, EEGLAB needs to be recompiled with the additional plugins.
- When using scripts, it is not possible to use external custom MATLAB functions or to define new functions
- When using scripts, the range of possible MATLAB commands is limited to the MATLAB core commands, all EEGLAB commands and a few commands from the signal processing, the statistics and the optimization toolbox (commands which are included in the compiled code).
- It is not possible to modify EEGLAB source code.
Can I run some of my MATLAB scripts on the EEGLAB compiled version? Yes, as long as they use standard MATLAB functions or EEGLAB functions.
- Is the compiled code faster than the non compiled one? No, it is the same speed as the MATLAB runtime engine still interprets it.
- Is it legal since I do not own MATLAB? Yes, it is perfectly legal. Although it is illegal to run a pirated version of MATLAB, it is perfectly legal to distribute the MATLAB Runtime Engine and compiled MATLAB code.
- Can I use this program for commercial applications such as my private clinical practice? EEGLAB is a research software package, not a clinical software package. The program is provided as-is with no warranty of any kind. Although the EEGLAB core code is released under a BSD 2.0 license that allows for commercial use, some of the plugins included in the compiled version are not (this includes plugins handling source reconstruction, for example).
- Trouble shooting! Do not hesitate to submit bug reports on the EEGLAB Gihub issue tracker if you encounter a problem running a compiled version of EEGLAB.
Differences between the MATLAB and the compiled EEGLAB version mostly arise because of the way external data files are handled. This is how we check the compiled version’s integrity (since it is not possible to automate this process yet). You may run the script test_compiled_version.m in the folder functions/support_files (you might have to change the path to the data files). For testing, we use the STERN STUDY (0.9 Gb). Please download the data on your computer.
- Start EEGLAB
- Edit options and change them; go back to option and make sure they have changed
- Load the continuous tutorial dataset “eeglab_data.set”. Look up electrodes, run clean raw data, Run ICA, run IClabel
- Perform source localization of ICA components using Dipfit
- Plot 3-D ERP after coregistering
- Load Stern study, precompute ERPs and plot them. Try using statistics.
- Check that the help menus are functional
If all of the above checks, then the compiled EEGLAB version is considered verified. We check the version separately on Windows and Unix/MacOS.
Compiling EEGLAB usually consists of opening the “eeglab.prj” file (in the root EEGLAB folder) using the MATLAB Compiler Graphical App, and pressing the package button. However, there might be some path issues that require fixing on your system. See detailed instructions below.
- Clone EEGLAB with default plugins
git clone --recurse-submodules https://github.com/sccn/eeglab.git
- Install additional plugins (plugins that are already installed will be skipped). Some folders from the plugins clean_rawdata and Fieldtrip should be removed to avoid compilation issues. Use the following script to install plugins and remove these folders:
eeglab % restart the fleshly installed eeglab % Installing plugins plugin_askinstall('ANTeepimport', 'eegplugin_eepimport', true); plugin_askinstall('Fieldtrip-lite', 'ft_defaults', true); plugin_askinstall('Fileio', 'ft_read_data', true); plugin_askinstall('IClabel', 'eegplugin_iclabel', true); plugin_askinstall('PICARD', 'picard', true); plugin_askinstall('vised', 'vised', true); plugin_askinstall('bids-matlab-tools', 'bids_export', true); plugin_askinstall('bids-validator', 'pop_validatebids', true); plugin_askinstall('bva-io', 'eegplugin_bva_io', true); plugin_askinstall('clean_rawdata', 'eegplugin_clean_rawdata', true); plugin_askinstall('dipfit', 'eegplugin_dipfit', true); plugin_askinstall('egilegacy', 'eegplugin_egilegacy', true); plugin_askinstall('firfilt', 'eegplugin_firfilt', true); plugin_askinstall('irrfilt', 'eegplugin_iirfilt', true); plugin_askinstall('musedirect', 'eegplugin_musedirect', true); plugin_askinstall('musemonitor', 'eegplugin_musemonitor', true); plugin_askinstall('neuroscanio', 'eegplugin_neuroscanio', true); plugin_askinstall('xdfimport', 'eegplugin_xdfimport', true); % Removing clean_rawdata files % For clean_rawdata, remove folder manopt/reference/m2html. CleanRawData_folder = fileparts(which('clean_rawdata.m')); rmdir(fullfile(CleanRawData_folder,'manopt','reference','m2html'), 's'); % Removing FieldTrip files % For Fieldtrip remove folders compat, external/afni, external/spm8, external/spm12, external/gifti, external/eeglab, external/bemcp and external/npmk FieldTrip_folder = fileparts(which('ft_defaults.m')); rmdir(fullfile(FieldTrip_folder,'compat'), 's'); rmdir(fullfile(FieldTrip_folder,'external','afni'), 's'); rmdir(fullfile(FieldTrip_folder,'external','spm8'), 's'); rmdir(fullfile(FieldTrip_folder,'external','spm12'), 's'); rmdir(fullfile(FieldTrip_folder,'external','gifti'), 's'); rmdir(fullfile(FieldTrip_folder,'external','eeglab'), 's'); rmdir(fullfile(FieldTrip_folder,'external','bemcp'), 's'); rmdir(fullfile(FieldTrip_folder,'external','npmk'), 's'); rmdir(fullfile(FieldTrip_folder,'external','signal'), 's'); rmdir(fullfile(FieldTrip_folder,'external','mffmatlabio'), 's'); rmdir(fullfile(FieldTrip_folder,'external','egi_mff_v2'), 's');
Optional: Edit the eeglab.m file and add new plugins to the compiled version of EEGLAB (line 900).
Open the “eeglab.prj” file in the Matlab editor. Check the path for plugins. If a new version is available, rename the version in the eeglab.prj file.
Change the EEGLAB version in the eeg_getversion.m and in the eeglab.prj files.
Open the Application compiler (Matlab tab “Apps” and button “Application compiler”) and open the “eeglab.prj” file. DO NOT RESAVE THE PROJECT IN THE APPLICATION COMPILER AS IT TENDS TO MESS UP PATHS FOR CROSS PLATFORM COMPILATION.
On the command line, type “setenv(‘MCC_USE_DEPFUN’,’1’)”. There will still be class errors when checking dependencies, but the EEGLAB will compile anyway.
Press “Package” and wait (usually 30 minutes or so)
If successful, 3 folders are created. You may test the compiled EEGLAB version by running the program in the “for_testing” folder.
Test the compiled version for potential runtime errors (see notes on testing here). On Mac and OSX use ./run_eeglab.sh MATLAB_PATH.
Known error: running eLoreta from DIPFIT, cannot find precompute_dpss
Selected for future inclusion but require testing. If you have a plugin you want to include, please try compiling with the plugin and testing the plugin in the compiled EEGLAB version. If all is functional, email us at eeglab_at_sccn.ucsd.edu and your plugin will be included in the next release.
Adding new plugins
- Download or clone plugin
- Add to eeglab.m
- Update this page (create a pull request)
|Biosig||Path issue for compiler. Using the version included in Fieldtrip|
|MFFMatlabIO||Issue with finding the JAR file at execution time; more debugging necessary before inclusion possible|