Automatic EEG IC (ICA) Classification for brainstorm

,

Hi everyone,

I saw there was a similar made in Feb 2018 by @alirezatalesh that received no response.

EEGLAB offers ICLabel plugin that can be applied to ICA for automatic component detection. It gives an estimation of probable artefact or lack of artefact. The latter is recognised as brain activity.

I've had a good look at the topics posted and the tutorials available on brainstorm, but I haven't seen anything similar offering automatic IC classification for EEG data, available to use on brainstorm. I am wondering if there has been such or similar plugin introduced for brainstorm since the last post was made?

I have also attached a screenshot of the EEGLAB output.

Thank you,
Vyte

Screenshot 2022-03-02 at 12.19.19

No, there is nothing as this nice automatic classification. This is something I would love to see in Brainstorm, but unfortunately we lack resources for developing all the tools we are interested in...
I wonder how easy it would be to connect this automatic labeling in Brainstorm. Since we are already using the EEGLAB runica.m, there are probably ways to do this...
I hope I'll be able to work on this at some point this year.

There is one option that might be of interest to you: the text box " Sort components based on correlation". Description: https://neuroimage.usc.edu/brainstorm/Tutorials/Epilepsy#Artifact_cleaning_with_ICA

Sort ICA components based on their correlation with a selection of 'reference' channels (default: ECG & EOG). The sorting is based on the maximum correlation of a given ICA time-series with any of the reference channels.

1 Like

Hi Francois,
Thank you for your response.
I did anticipate this. Automatic classification would indeed be a useful tool in Brainstorm and I wish you the very best on this future project!

In terms of correlation, I already have a script written for extracting and correlating ICA time-series components with ECG. It was my intention to use the ICLabel option to guide me and confirm those suspicions using correlations.

Thank you for your help!
Vyte