In the documentation for find_events, there is the following warning:
If you are working with downsampled data, events computed before decimation are no longer valid. Please recompute your events after decimation, but note this reduces the precision of event timing.
How can I recompute events after decimation, given that decimation is performed by mne during epoching?
Hello @axel,
the warning is really only relevant if you decimate the raw data after you’ve extracted the events. If you’re only passing decim to mne.Epochs() (and not decimating the raw data before), you don’t have to worry about the events being shifted or timing precision being lost – at least that is my current understanding!
Best wishes,
Richard
Events are stored as sample numbers, which means they inherently depend on the sampling frequency. Therefore, if you change the sampling frequency, you also need to adjust events. This does not apply for annotations though, because these are time-based (in seconds).