After running autoreject, the pipeline should be able to write a new EDF based on the results from autoreject, for which multiple steps must be implemented:
[ ] When an epoch is removed based on autoreject, all of the annotations from this point on should be shifted by the $\Delta t$ removed (i.e., the number of samples from the removed epoch). This must be performed in a cumulative approach for all the removed epochs.
[ ] Investigate what happens when autoreject rejects just one channel in a specific epoch so this can also be taken in consideration.
After running autoreject, the pipeline should be able to write a new EDF based on the results from autoreject, for which multiple steps must be implemented: