Band-pass filtering after ICA

Discussions
,
1

Dear Francois!

I noticed, that by mistake I removed ICA projectors before applying band-pass filter to data. What I did, was that I used 0.5Hz - 100Hz frequency band for display filters, ICA calculations (continuous data) and finally for the band-pass filtering of the data. In band-pass filtering I selected the option "Process the entire file at once".

I know that doing like this is something that you do not recommend. In tutorial: https://neuroimage.usc.edu/brainstorm/Tutorials/ArtifactsFilter#Apply_a_notch_filter
"When to apply these filters?
*Before SSP/ICA cleaning: Artifact cleaning with SSP/ICA projectors require all the channels of data to be loaded in memory at the same time. Applying a frequency filter on a file that contains projectors requires all the file to be loaded and processed at once, which may cause memory issues. Pre-processing filters are rarely changed, whereas you may want to redo the SSP/ICA cleaning multiple times. Therefore it is more convenient to apply the filters first."

I understood, that I could have had problems with computer memory, but that my data should still be valid. Is this true?

As usually, I noticed my mistake after I had all the results ready. So if you happen to know the answer, I would really appreciate your help.

Thank you in advance!

2

There is no specific reason for the ICA not to work on broadband signals, or issues for filtering the signals from which a few ICs were removed.
However, the results can very different. The ICA of the broadband signals may include many components that do not exist in the bandpass-filtered signals, and therefore produce a completely different decomposition.

The more important is probably that you use the exact same procedure for all the subjects, and that you apply the same pre-processing steps to the recordings with which you estimate the noise covariance matrix.