[Eeglablist] Preprocessing suggestions

Stefan Debener stefan.debener at uni-oldenburg.de
Wed May 19 00:38:09 PDT 2021

Hi Chiara,

It is difficult to give precise recommendations. How "noisy" EEG data 
are during physical activity such as cycling will depend on the 
recording conditions, the hardware used, and the participants, of 
course. Additional information like cadence event markers, power output, 
IMUs, or video footage during data acquisition would help guiding the 
analysis. Moreover, it matters whether you have stationary cycling or 
real cycling data, whether wireless EEG with a head-mounted amplifier 
was used (hopefully), what kind of cap and electrodes, etc. In any case, 
ICA benefits from detrending (approx. 1 Hz high-pass) and pruning (try 
to avoid manual pruning) before decomposing the data. Moreover, it does 
not make too much sense to me to determine the number of components to 
be removed beforehand - but it makes sense to use semi-automatic tools 
for component selection (depending on your experience). to get started, 
I suggest you check out Makotos pre-processing pipeline...

In order to minimize the risk of circular analysis 
(https://urldefense.com/v3/__https://pubmed.ncbi.nlm.nih.gov/19396166/__;!!Mih3wA!UWFLW_EAAfrgahU2qSH_ZvoiHLvG4cpP1yIyj_XhMGAr88OuypST0Q2F2wd2pipeOMMS6A$ ) you may want to take a few 
pilot datasets, develop your optimal processing pipeline on these few 
datasets only and then apply this pipeline to all other datasets.

Hope this helps.



On 18.05.21 16:27, Chiara Gattoni via eeglablist wrote:
> ACHTUNG! Diese E-Mail erreicht Sie von einem Absender außerhalb der Universität Oldenburg. Bitte klicken Sie auf keine Links oder öffnen Sie keine E-Mail-Anhänge, falls Sie den Absender nicht kennen und nicht wissen, ob der Inhalt sicher ist.
> WARNING! This email originated off-campus. Please do not click on any links or open any email attachments if you do not know the sender and are unsure as to whether the content is safe.
> Hi all,
> I am in the process of starting the preprocessing of my EEG data which
> have been collected during a cycling task of 40 minutes duration under
> different conditions. Participants were cycling at a constant intensity
> (moderate intensity domain).
> This is my first time analysing EEG during physical activity and one of the
> first times analysing EEG data in general (I come from sport and exercise
> sciences).
> I would like to do frequency and time-frequency analyses and focus on
> alpha, beta and theta bands. No events have been measured and I was
> thinking of dividing the 40-min EEG data collected into 5-minute time
> intervals.
> Have you got any suggestions to give me regarding the preprocessing part?
> Data collected are going to be more noisy than resting EEG for sure, so if
> you have any specific tips it would be helpful.
> These are the steps I was thinking to follow:
> 1) Baseline removal;
> 2) Bandpass filter (0.5 - 30 Hz);
> 3) Downsampling (128 Hz; sampling rate was 500 Hz);
> 4) Removing artifacts manually;
> 5) ICA;
> 6) Using SASICA for removing bad components (2 maximum);
> 7) Epoching (4 seconds overlap).
> Would you please let me know if I am going in the right direction?
> Many thanks in advance,
> Chiara
> _______________________________________________
> 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

Prof. Dr. Stefan Debener
Department of Psychology
Neuropsychology	Lab
University of Oldenburg
D-26111 Oldenburg

Office: A7 0-038
Phone: +49-441-798-4271
Fax:   +49-441-798-5522
Email: stefan.debener at uni-oldenburg.de

More information about the eeglablist mailing list