[Eeglablist] Methods question: Resting EEG and machine learning

[Delorme, Arnaud]

Dear Rachel,

Attached is a good review on deep learning (I think the same conclusion would apply to regular machine learning) and EEG. About 40% of the models train models on the raw data and the rest on extracted features obtained from the raw data. Using the raw data is thus common although my intuition is that you would need more data so the model can “learn” to extract features from it.

https://urldefense.com/v3/__https://iopscience.iop.org/article/10.1088/1741-2552/ab0ab5__;!!Mih3wA!UIs-hhNNknupFaXD3LdK0m8AqhodZMNYwVqlSshXKYwlpMx84XIP_wTUaorB9UlxWY67ew$ 

Best wishes,

Arno

> On Jul 17, 2020, at 1:19 PM, Makoto Miyakoshi <mmiyakoshi at ucsd.edu> wrote:
> 
> Dear Rachel,
> 
> Just a piece of general advice--scalp electrode potentials depends on the
> choice of reference. You want to make sure which reference you are using,
> and the previous studies you are referring to. Otherwise I can't think of
> anything specific off the top of my head.
> 
> Makoto
> 
> On Fri, Jul 10, 2020 at 10:00 AM Judy,Rachel L <rachel.judy at ufl.edu> wrote:
> 
>> Hello all,
>> 
>> Our lab has resting EEG data (10 mins of continuous data) collected from
>> participants across 3 different time points.
>> After preprocessing and cleaning we are feeding data into a machine
>> learning algorithm. The data are in electrode space and we are feeding the
>> learner an average voltage across time for each electrode. We have some
>> concerns with this approach but we found this paper by Yildirim 2019
>> https://urldefense.com/v3/__https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6406978/pdf/ijerph-16-00599.pdf__;!!Mih3wA!Tjzk5yJ1J_B2bnKc5KaAQQFE2Sswl0OC9srjmcesYbrNUYwTp0pg5OBg27sEu1gpVi3yUw$
>> that uses a similar technique but we remain cautious given issues with
>> meaningful interpretation of a potential difference in resting voltage
>> between subjects and between different data collection sessions. Does
>> anyone have any suggestions or thoughts on the validity and reliability of
>> using voltages in this manner?
>> 
>> Many thanks
>> 
>> 
>> -Rachel
>> 
>> 
>> 
>> _______________________________________________
>> 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
>> 
> _______________________________________________
> 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