[Eeglablist] preprocessing continuous data for spectral analysis with welch approach

[Makoto Miyakoshi]

Dear Agnieszka,

First of all, if your data has poor quality you can't expect good results
even if you perform the fansiest signal processing. This is called GIGO
(garbage-in, garbage-out). There is no established way to evaluate the EEG
data quality in a non-relative sense (i.e. without depending on
experience). It's better to ask someone experienced to evaluate your data
quality.

For cleaning continuous data, I recommend using ASR in clean_rawdata()
plugin as a pre-ICA processing, then ICA. If this approach does not work, I
would doubt the data quality.

If you are interested in my personal memos for preprocessing pipeline, see
this page.

http://sccn.ucsd.edu/wiki/Makoto's_preprocessing_pipeline

Makoto



On Mon, Jun 13, 2016 at 7:30 AM, Agnieszka Zuberer <azuberer at googlemail.com>
wrote:

> hi everyone,
>
> we would like to perform a spectral analysis for Theta, Alpha and Beta
> power based on the welch approach on our continuous data sets, recorded
> with 3 scalp electrodes (FZ, CZ, OZ) referenced to the mastoid.
>
> We gained bad results with both regression based artifact corrections
> (Gratton, 1983) and ICA. Reading the preprocessing tutorials of eeglab for
> continuous datasets, we wonder about the order of following procedures
>
> a) reject continuous portions of data based on spectrum (with pop_rejcont)
> b) reject continuous portions by simple mv-threshold (iterating through
> the raw data in 1 second-steps and reject portions exceeding 150mv)
> c) perform pwelch spectral analysis
>
> Our question is, what is worse:
> - performing a pwelch approach on "cleaned" data resulting in breakes in
> the data
> *or *
> - performing a pwelch approach on raw uncorrected data while maintaining
> the continuity of the data and reject data afterwards with indices for
> segments marked for rejection
>
> thank you very much for helping us out.
>
> Agnieszka
>
>
>
> _______________________________________________
> 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
>



-- 
Makoto Miyakoshi
Swartz Center for Computational Neuroscience
Institute for Neural Computation, University of California San Diego
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160701/00ae8ee3/attachment.html>