[Mne_analysis] Concatenate preprocessed epochs

[Dirk van Moorselaar]

Hello,

I have a follow up question on a question that I asked previously.
In my experiment I have collected data into seperate sessions, on separate days. I apply preprocessing (artefact detection, Ica, bad electrode selection) on these sessions separately and after preprocessing and spline interpolation to correct bad channels I would like to combine these separate sessions int a single -epo.fif file. 

To do this I use mne.concatenate_epochs(). However, this fails and gives me an assertion error. I understood from my previous question that to get this to work  I need to remove the bad channels from both sessions (even if we are dealing with separate channels in the two sessions) and then use interpolation on the concatenated data before further processing.

At first sight I do not necessarily like this solution as in some cases I have to remove perfectly fine channels from one session, even though their signal was good  in that session. Is there a workaround that still allows you to concatenate epochs, but does not force you to have the same channels marked as bad in both sessions? 

Thank you again,

Dirk