[Eeglablist] select baseline range based on epoch window

[Tarik S Bel-Bahar]

Hello Yadwinder ,  brief notes below, best wishes.





************Notes for Yadwinder

This topic has been discussed several times on eeglablist, try googling it
eegalblist + your topic
regarding both baselining and epoch lengths, if you are just beginning your
best bet is follow the methods from high-quality articles that use EEG or
MEG and a paradigm similar to yours , and have epochs that are as long as
the ones you have chosen to use.

Other resource to check if you haven't already include the tutorials and
settings from any EEG software (e.g., eeglab, fieldtrip, brainstorm, mne
ptyhon, etc) . EEG handbooks also have discussions of usual baselin i ng
and epoching methods/choices.

Generally, eeglab prefers common length epochs across single-subjects.
Recent versions of eegalb study may allow you to work with varying length
epochs.

Overall your question depends on whether your protocol is event-related or
continuous.
48sec is a long time compared to usual epochs in EEG research (which are
usually ~1 to 3 seconds long, and event-related). It may be best for you to
have a way to cut these longer periods into smaller periods, perhaps
related to various events during the 48 second period.
It is also unlikely that these 48s periods are without artifacts such as
blinks or movement, so double check to see how you have cleaned your data,
and whether you truly have many 48 sec that are clean of artifact.

Also, if you are going to stay with the 48s periods, it's important that
you have many epochs like this for each participant (~20 or 30 minimum).

For sanity checks and your initial review of the data patterns, do an
initial analysis using the shortest length epoch and use that as a length
for all epochs.

Alternatively, make subclasses of epochs that are at different lengths
(e.g., the 3, 5, and 10 s ones)

Alternatively, look at event-related activity time-locked NOT to the onset
of the stimulis, but to the "response" of the participant, or to some other
event.

regarding baselining, baselines in the literature range from -1000 ms to
-100 ms, it depends on the paradigms. It also depends if you are looking at
ERPs or ERSPs, or if the data is continuous. It also depends on the quality
of the baseline period (is it clean, it is neutral, is it a valid baseline
period relative to the task?)

Note that one can also baseline to the mean of the whole epoch, to a
baseline gathered in a different or resting task, to an estimate across
all-subjects, etc.. Check previous eeglablist discussions on this, some of
which I've pasted below from a google search.



***** Below some examples from google scholar discussing baselining effects
Tanner, D., Norton, J. J., Morgan-Short, K., & Luck, S. J. (2016). On
high-pass filter artifacts (they’re real) and baseline correction (it’sa
good idea) in ERP/ERMF analysis. Journal of neuroscience methods, 266,
166-170.

Schr, B. M. E., & Widmann, A. (2016). High-pass filters and baseline
correction in M/EEG. cognition, 52(8), 997-1009.

Hu, L., Xiao, P., Zhang, Z. G., Mouraux, A., & Iannetti, G. D. (2014).
Single-trial time–frequency analysis of electrocortical signals: Baseline
correction and beyond. Neuroimage, 84, 876-887.



***below examples of google search output or eeglablist + baseline

Feb 25, 2017 - I usually baseline in the code and always make sure to save
the baselined ... Jumana <jumana.ahmad at kcl.ac.uk> wrote: > Dear EEGlab
list, ...

[Eeglablist] ERSP baseline

https://sccn.ucsd.edu/pipermail/eeglablist/2006/001332.html

Mar 16, 2006 - If my AD rate is 1000Hz, and my >baseline period is -1 to 0,
will the last ... >eeglablistmailing list eeglablist at sccn.ucsd.edu
>Eeglablist page: ...

[Eeglablist] Baseline

https://sccn.ucsd.edu/pipermail/eeglablist/2016/010906.html

Feb 5, 2016 - [Eeglablist] Baseline ... You can > also see what the usual
baselines are in their tutorial data. ... More information about the
eeglablist mailing list.

[Eeglablist] baseline correction

https://sccn.ucsd.edu/pipermail/eeglablist/2018/013941.html

Aug 15, 2018 - Previous message: [Eeglablist] baseline correction; Next
message: [Eeglablist] Batch ... say [-10 +10] from stimulus onset - that
would be your baseline, and taking the average ... More information about
the eeglablist mailing list.

[Eeglablist] Continuous data - baseline correction

https://sccn.ucsd.edu/pipermail/eeglablist/2016/010890.html

Jan 31, 2016 - [Eeglablist] Continuous data - baseline correction ... On
Mon, Jan 25, 2016 at 4:57 PM, Liz Chrastil <chrastil at bu.edu> wrote: >
Hello List ...

[Eeglablist] baseline removal and ICA - Swartz Center for ...

https://sccn.ucsd.edu/pipermail/eeglablist/2012/005003.html

May 30, 2012 - [Eeglablist] baseline removal and ICA ... at sccn.ucsd.edu"
<eeglablist at sccn.ucsd.edu> Enviado: Martes, 29 de mayo, 2012 4:53 P.M.
Asunto: Re: [Eeglablist] baseline ... More information about the eeglablist
mailing list.

[Eeglablist] When to select Additive Baseline in ERSP Study ...

https://sccn.ucsd.edu/pipermail/eeglablist/2018/013820.html

Jul 18, 2018 - [Eeglablist] When to select Additive Baseline in ERSP Study
Function? ... Van Benthem < kathy_vanbenthem at carleton.ca> wrote: > Dear
List, ...

[Eeglablist] extract epochs without removing baseline

https://sccn.ucsd.edu/pipermail/eeglablist/2012/005964.html

Dec 13, 2012 - [Eeglablist] extract epochs without removing baseline.
Stephen ... Apologies if this has been answered before, but I could not
find it in the > list.

[Eeglablist] time/frequency baseline length - Swartz Center for ...

https://sccn.ucsd.edu/pipermail/eeglablist/2006/001573.html

Oct 27, 2006 - Scott Makeig From: "Ayelet Landau" <ayeletlandau at
berkeley.edu> To: "EEGLab List" <eeglablist at sccn.ucsd.edu> Date: Thu, 26
Oct 2006 ...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20180920/97d08925/attachment.html>