Simply EEGLAB
Running ICA
Station 4 · Lesson 3 · EEGLAB 2026.1
In the GUI
- Datasets
- ToolsFilter the dataBasic FIR filter (new, default)
- Lower edge of the frequency pass band (Hz)OK
1 - Overwrite it in memory (set=yes; unset=create a new dataset)
D1 ica1hz - ToolsDecompose data by ICA
- ICA algorithm to use (click to select)Extended Infomax (runica.m; default)
'extended', 1 - Reorder components by variance (if that's not already the case)Use only channel type(s) or indicesOK
- Proposed rank:OK
- PlotComponent mapsIn 2-D
- Datasets
- EditDataset info
- ICA weights array or text/binary file (if any):from other datasetOK
2ALLEEG(2).icaweightsALLEEG(2).icasphereALLEEG(2).icachansind - OK
- FileSave current dataset as
Script
EEG_main = EEG;
EEG_ica = pop_eegfiltnew(EEG_main, 'locutoff', 1);
EEG_ica.setname = 'D1 ica1hz';
expectedRank = EEG_ica.nbchan - numel(badChans) - 1;
fprintf('Expected data rank: %d\n', expectedRank);
EEG_ica = pop_runica(EEG_ica, 'icatype', 'runica', 'extended', 1);
EEG_main.icaweights = EEG_ica.icaweights;
EEG_main.icasphere = EEG_ica.icasphere;
EEG_main.icachansind = EEG_ica.icachansind;
EEG_main.icawinv = []; % recomputed by eeg_checkset
EEG_main.icaact = [];
EEG = eeg_checkset(EEG_main);
pop_topoplot(EEG, 0, 1:20);
EEG = pop_saveset(EEG, 'filename', 'D1_main_ica.set', 'filepath', pwd);Which value should I choose?
| Setting | Value |
|---|---|
| High-pass for the ICA copy | 1 Hz, or 2 if components stay ambiguous |
| Algorithm | Extended Infomax (runica) |
| Channels | All EEG channels. In D1 we keep EOG1 and EOG2 |
| Number of components | Equal to the data rank: 31 in D1 after average reference |
Check
- In the main EEGLAB window
- In the command window
- Open the component maps
- Reload the saved file
- The rank matches
For your Methods
Independent component analysis (extended Infomax; Lee et al., 1999) was computed with runica ([31] components after PCA reduction to the data rank) on a copy of the data high-pass filtered at 1 Hz, and the resulting weights were applied to the 0.1 Hz high-pass filtered data.