Recap
- Referencing and filtering
- Removing unwanted channels
- Removing artefacts
Independent Component Analysis (ICA)
- Why we need ICA?
- ICA is a signal processing method to separate independent sources linearly mixed in several sensors (ICA for Dummies - By Delorme)
- Let’s watch the video together: ICA by Delorme
- Remove artefacts, make sure you have enough data (20 * square of number of channels), high pass filter at 0.5, remove bad channels.
Running ICA
- Use the the
Lab 4 demo
dataset on Learn.
- Use
Tools -> Decompose data by ICA
. Select runica
, extended 1
(helps to detect line noise).
- You can also plot all or individual component maps from the
Plot
menu.
- Use
Tools -> Inspect/label components by map
.
- Use
Tools -> Remove components from data
.
Continued
- Or, do this automatically:
Tools -> Classify components using ICLabel
, then flag components as artefacts.
- Important: DON’T remove/reject components if you are conducting group level analysis.
To submit
- Use the the
Lab 4 exercise
dataset on Learn.
- Conduct pipeline processing as per usual.
- Run ICA, share a screenshot.
- Remove components, share a screenshot.
- Compare pre- and post-analysis attributes.
- Compare pre- and post-analysis Channel data (scroll). What do you conclude? Any problems/issues? Why or why not? (consists 25% of today’s mark).