EEGLAB Documentation

The Resampling Statistical Toolkit is a package within EEGLAB designed for inferential statistics using resampling methods (bootstrap and permutation). Data are organized into arrays, allowing multiple tests to be executed simultaneously across dimensions such as time, electrodes, and subjects.

Requirements

• Matlab: Required for all functions.
• Statistics Matlab Toolbox: Required only if performing parametric statistics. It is NOT required for resampling methods (bootstrap/permutation).

Data Organization

To run multiple tests at once, organize your data into arrays where the last dimension represents the subjects. For example, a design with 100 time points, 64 electrodes, and 10 subjects should be shaped as (100, 64, 10).

The EEGLAB repository is organized into several key directories:

• /functions: Contains all distributed EEGLAB functions, categorized into admin, sigproc, pop, and misc.
• /plugins: The location where downloaded EEGLAB plug-ins should be placed. dipfit (1.0) is included by default.
• /sample_data: Contains miscellaneous EEGLAB data used for tutorials and references.
• /sample_locs: Contains standard channel location files (e.g., 10-10, 10-20).

To start the EEGLAB environment, follow these steps:

1. Launch Matlab.
2. Use the Matlab file browser or cd command to navigate to the directory containing the EEGLAB installation.
3. At the Matlab command prompt (>>), type eeglab and press Enter.

Note: If you encounter permission or execution issues, try running Matlab as an administrator.

>> eeglab

To install EEGLAB using Git, you must use the --recurse-submodules flag. Downloading a ZIP file directly from GitHub is not recommended because it will not include the necessary EEGLAB submodules required for full functionality.

If you have already cloned the repository without submodules, you can initialize them by navigating to the eeglab folder and running the submodule update commands.

# Recommended way to clone
git clone --recurse-submodules https://github.com/sccn/eeglab.git

# If you forgot to include submodules during the initial clone
git submodule update --init --recursive --remote
git pull --recurse-submodules

After running statistical tests (like statcond), you can use the fdr function to implement the False Detection Rate method to correct for multiple comparisons.

p = fdr(p);

For non-trivial layouts, use the 'geom' input in inputgui. This works similarly to the Matlab subplot function, allowing you to define specific bounding boxes for each control.

Each geometry entry is a cell array: {rows, cols, [x_start, y_start], [x_end, y_end]}.

res = inputgui('geom', { {2 1 [0 0] [1 1]} {2 1 [1 0] [1 1]} }, 'uilist', ...
                          { { 'style' 'text' 'string' 'Enter a value' } ...
                            { 'style' 'edit' 'string' '' } });

EEGLAB provides enhanced versions of standard Matlab dialog functions that use the inputgui toolkit for a consistent look and feel. Use these instead of the built-in Matlab functions:

• questdlg2: Replaces questdlg (Question dialog)
• errordlg2: Replaces errordlg (Error dialog)
• warndlg2: Replaces warndlg (Warning dialog)
• listdlg2: Replaces listdlg (List selection dialog)
• inputdlg2: Replaces inputdlg (Input dialog)

% Example using questdlg2
ButtonName = questdlg2('What is your favorite color?', 'Color Question', ...
                         'Red', 'Green', 'Blue', 'Green');

The following functions are available in the Resampling Statistical Toolkit: