Should the units of `OptogeneticSeries` be watts per unit of area instead of just watts?

[h-mayorquin]

Currently, the units of OptogeneticSeries are fixed to watts:

https://nwb-schema.readthedocs.io/en/latest/format.html#optogeneticseries

I was discussing this today with a member of the Cohen lab, who mentioned that this is insufficient from a provenance standpoint. To quantify the intensity of the stimulus, it is necessary to understand how the wattage is distributed over a given area. They do this by measuring the intensity in a sensor and then dividing the value read by the sensor's area.

Has this aspect already been implicitly assumed or considered previously?

[h-mayorquin]

Alessandra brought up an important distinction is wattage at the source vs wattage at the stimuli site. The former is easy to fetch from the experimentalist whereas the second one, as mentioned above, would require an extra measurement. Ben suggested that we should just make the units flexible. Another option is two have two distinct attributes.

[h-mayorquin]

Relevant: https://github.com/rly/ndx-optogenetics

Where it is written as two different sources but this is not a TimeSeries. It is only a static source.

[rly]

I think it is fine to make the units flexible.