Timing for referencing EEG preprocessing

Dear community,

I am in the middle of writing all scripts for the preprocessing, ICA, epoching and analysis of my EEG data, using eeglab/matlab.

According to Lucks pipeline (which I am basing myself on; Applied Event related potential data analysis, 2022) we shouldn't reference the data at the start (when importing the data), but rather after filtering or, in case of referencing to an average of mastoids, after ICA. I was wondering if anyone knew the rationale behind this idea and whether this is significantly changing anything or not. Thanks in advance!

Kind regards

Helen Tobback

PhD student VUB, Belgium

Answer 1

[Taylor]

Here is a detailed explanation on the biological/electronics theory behind referencing for EEG.

https://eeglab.org/tutorials/ConceptsGuide/rereferencing_background.html