[Eeglablist] EEGLAB: Epoching and ICA. In which order?

[Bruzadin Nunes, Ugo]

Hi Alice,

I'm no expert, but here at the INL we've experimented with many orders of cleaning process. We've settled in cleaning the data continuously with ICA first, epoching later. This decision was made due to the fact that our analysis depends on epoching more than in components, so we use ICA to clean the data and epoch removal for a later more fat-trimming cleaning process (only delete really bad epochs). ICAs (PCA) cleaning can assist on deleting temporal-related problem-components such as ECG and chronic EMG and blinks, that get lost or mixed-in once the data is epoched. However, once PCA'ed, the data is now limited in rank which may limit later analysis.

In any case, we make sure that our data is averaged before epoch cleaning so we only remove high decibel errors and not straight channels near the reference, which happened quite a few times. Also, continuous ICA's are really slow, so you must have computer power or time (or both) to perform them.

But if your analysis will be component-focused, epoching and cleaning first might be more useful. It really is a question of what analysis will be made later and what data is of higher importance. ERP analysis for instance will average the errors, so gross ICA cleaning is less of an issue, epoching first might make more sense.

Hope it helps,

Best of luck,

Ugo Bruzadin Nunes, M.A. & Doctoral Candidate
PSYC 222 Instructor - Effects of Recreational Drugs
Brain and Cognitive Sciences Ph.D Program ||
School of Psychological and Behavioral Sciences
Southern Illinois University - Carbondale

________________________________
From: Xiaonan L. Liu <liuxiaonan87 at gmail.com>
Sent: Tuesday, September 10, 2019 2:40 PM
To: Alice Grazia <alicegrazia94 at gmail.com>
Cc: eeglablist at sccn.ucsd.edu <eeglablist at sccn.ucsd.edu>
Subject: Re: [Eeglablist] EEGLAB: Epoching and ICA. In which order?

Hi Alice,

It seems to me there's no absolute right or wrong protocol for this . I
usually prefer doing epoch before ICA and the reason is by epoch, you can
get rid of noisy portions of your data (e.g. at the beginning, breaks and
the end etc). It seems to me what's really important is you need clean data
in order to get good ICA decomposition and epoch is somewhat a way to get
cleaner data.

Best,
Xiaonan

On Tue, Sep 10, 2019 at 9:53 AM Alice Grazia <alicegrazia94 at gmail.com>
wrote:

> Hi all,
> I have this doubt that I could not solve by myself. Maybe it is more a
> theoretical question than a programming question, but I'm performing
> time-frequency analysis with EEGlab software for my master thesis and I'm
> not sure whether it is more correct to perform ICA and epoching in a
> precise order. I found some recommendations (EEGLAB tutorial 2004) saying
> that it is advisable to perform ICA on a greater amount of data as
> possible. However, if I run ICA on EEGLAB before epoching, I cannot have a
> good visualisation of the ICA components (only the maps and components
> scroll). Instead, I have seen in a tutorial by Mike Cohen, that he performs
> epoching before ICA, so there is the possiblity to look at each component
> latency and frequency before rejecting. Personally, I found myself better
> with the second method, but I am not sure wheather it is correct or not.
> I hope you can help me in this dilemma! Thank you!
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu
>