[Eeglablist] Issue with the pop_subcomp function

[Arnd Meiser]

Dear community,

I am currently running an ICA on my pre-processed EEG data (band filter, removal of bad epochs) and try to apply the weights of the ICA to the raw data as is recommended. After using the function pop_subcomp and thereby removing the identified bad components (lateral eye-movement, heartbeat) from the data, the amplitude of the signal is suddenly unrealisticly high (roughly from the hundreds to the billions). It seems I can downscale the signal again to reasonable values, but I would like to understand what is happening here. Has anyone had this issue before? When I apply the weights to the processed data, everything works the way I would expect it to.
I am grateful for any pointers.

Thanks and all the best,

Arnd Meiser