A2. Options to Maximize Memory and Disk Space
A2.1. Maximize memory menu
This section is
intended for users who use large EEGLAB datasets and need to optimize their
use of main memory (RAM) and disk space. To modify the relevant EEGLAB options,
select File > Maximize memory. If
you cannot modify the file eeg_options.m in the toolbox
distribution, the window below will pop up.
Simply press Yes. A copy of eeg_options.m will be stored
in the local directory and modified when you change any of the EEGLAB options.
(Note: For EEGLAB to use this file, the
current directory (.) must appear BEFORE the EEGLAB toolbox directory in
the Matlab path; see path()). If the original eeg_option.m file is writable,
EEGLAB will modify it instead.
If the original distribution copy of the options file is
modified, the new options will affect all EEGLAB sessions using that file.
If, however, EEGLAB (or the user from the Matlab or shell commandline) copies
the options file to the current working directory, then only EEGLAB sessions
having this directory in their Matlab path (before the EEGLAB distribution
directory) will follow the modified options. It is advised to have only one such option file to avoid confusion. Now the following
window will pop up.
When the top option is set, EEGLAB can hold in memory
more than one dataset at a time. New datasets are
created by most of the operations called under the Tools menu. With this option set, users can
undo dataset changes immediately by going back to working with the parent
(previous) dataset (unless they set the "overwrite dataset" option when they saved the new dataset). Processing multiple datasets may be necessary when comparing
datsets from separate subjects or experimental conditions.
The second option allow to save the raw data in a separate file. This will be useful in future version of EEGLAB to read one channel at a time from disk. This also allow faster reading of dataset when they are part of a STUDY.
If the 3rd option is set, all the ICA activitations (time courses of activity)
will be pre-computed each time the ICA weights and sphere matrices are
specified or changed. This may nearly double the main memory (RAM) required
to work with the dataset. Otherwise, ICA activitations
will be computed by EEGLAB only as needed, e.g. each time
EEGLAB calls a function that requires one or more activations. In this case,
only the activations needed will be computed, thus using less main memory.
The bottom option is used to remember folder when reading datasets. Using a text
editor, you should edit file "icadefs.m" in the distribution before beginning
to use EEGLAB. This file contains EEGLAB constants used by several functions.
In this file, you may:
A2.2. The icadefs.m file
TUTORIAL_URL = 'http://sccn.ucsd.edu/eeglab/eeglab.html'; % online version
ICABINARY = 'ica_linux2.4'; % <=INSERT
name of ica executable for binica.m
You can also change the colors of the EEGLAB interface using other options
located in the icadefs file.
Data Structures
Dipole localization using DIPFIT
