[Eeglablist] Does pop_clean_rawdata depend on the type of high-pass filt (0.1 or 1 Hz)?

Velu Prabhakar Kumaravel velu.kumaravel at unitn.it
Tue Jul 27 03:00:30 PDT 2021


Hi Bianca,

I have observed minor differences in the list of bad channels if you use
the same data consequently. This is because *clean_rawdata *uses a robust
approach (RANSAC) to compute the Pearson correlation of channels with their
neighbors and every time you run the algorithm, it might consider a
different set of samples leading to slightly different resultant
correlation values (This is rare, though).

Filtering EEG at 1 Hz would reduce much more low-frequency noise peaks
compared to filtering at 0.1 Hz. This can change the list of bad channels,
especially for noisier subjects. This is more likely to be true for any
other bad channel detection method.

Given these, you may

1) apply 1 Hz filter to the raw data
2) detect bad channels using clean_rawdata (store the list of bad channels)
*         bad_channels = find(EEG.etc.clean_channel_mask == 0); % I used
this command in the past, please verify if it works*
3) perform ICA

4) apply 0.1 Hz filter to the raw data
5) remove *bad_channels *manually computed in step - 2
6) transfer ICA weights computed in step - 3


Hope this helps.

Best regards,

Velu Prabhakar Kumaravel, Ph.D. Student
Center for Mind/Brain Sciences,
University of Trento, Italy


On Mon, 26 Jul 2021 at 17:09, Bianca Monachesi via eeglablist <
eeglablist at sccn.ucsd.edu> wrote:

> Dear EEGlab users,
>
> I used "pop_clean_rawdata" to find and remove bad channels (after filtering
> step) but the number of channels found on a signal high-pass filtered at 1
> hz is different from the number of channels found on a signal filtered at
> 0.1 hz. Does it make sense? Which is the relationship between filtering and
> the function clean_rawdata?
>
> This unbalance in the number of channel can be insidious since according to
> Mokoto's preprocessing pipeline (
> https://sccn.ucsd.edu/wiki/Makoto%27s_preprocessing_pipeline), the ICA
> performed on a signal high-pass filtered at 1 Hz  should be attached on a
> signal high-pass filtered at 0.1 Hz, considering that in both signals the
> number of channels removed and interpolated is the same...
>
> Thank you in advance for any kind of help or explanation of the problem.
>
> Best,
>
> bm
>
>
>
> --
> Bianca Monachesi, PhD
> Department of Psychology and Cognitive Sciences
> University of Trento
> Corso Bettini 31, Rovereto, TN (Italy)
> _______________________________________________
> 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
>


More information about the eeglablist mailing list