Extracting session duration

[twheatcroft07]

I am interested in understanding causes of variability in good unit number (relative to total unit number, or numbers of channels). I have been using data such as this .

Is it possible to extract the duration of the extracellular electrophysiological recording? I can see the session start time, but not the end, and I am unsure whether this represents the duration of any acquisition (e.g. of cameras) or of the ephys specifically.

Thanks!

[CodyCBakerPhD]

There is no official stop time in NWB; the 'length' of the session can be inferred by iterating over all the main data streams and estimating it based on the maximum timestamp[-1] (last index of time vector) or alternatively time_series.shape[0] * rate (since the first axis of any series should be time) over all the temporal objects

[twheatcroft07]

Perfect, thanks!

[CodyCBakerPhD]

However I caution that for chronic studies, there may be many breaks in between the start and end of what constitutes a 'session', usually annotated by the epochs table