Reducing RAM USAGE

The ERPWAVELAB files including all epochs are very large and require a lot of RAM. Often ‘Out of Memory’ problems in MATLAB occurs if care is not taken as to how the files are generated and processed. Here is a list explaining how RAM usage in ERPWAVELAB can be minimized:

1) Make sure EEGLAB or other programs don’t take up any unnecessary RAM. NOTICE! It is not necessary to keep the raw data files in EEGLAB open when running ERPWAVELAB.

2) The full array including all epochs is only needed for ERPCOH analysis. If you are investigating other measures you don’t need the full array. Thus, select:
 file -> Create ERPWAVELAB file ->ERPWAVELAB (current measure)
to store a small file only containing the measure investigated. Load this file into ERPWAVELAB and remove the existing large file.

3) Select: Options -> Optimize Ram Usage and make sure this menu point is checked. The data used for processing will only be the data in the region of interest (ROI) not keeping in memory the channels removed and the time-frequency regions outside of the ROI.

4) Split the dataset in EEGLAB into datasets containing less epochs. Then analyze the total dataset by loading each file containing a subset of all epochs into ERPWAVELAB and select ‘Use All’.

5) Alternatively, create your ERPWAVELAB file using a smaller resolution.

Developed by Morten Mørup

A tOOLbox FOR MULTI-CHANNEL TIME-FREQUENCY ANALYSIS

Back to main help page