In sleep electrophysiology we have a situation where basically every lab is talking about the same set of phenomena (slow waves, sharp wave-ripples, spindles, etc), but detecting them with different detectors (e.g. from a different signal, filter/window parameters, etc).
I used to think this was an issue for replicability: how can you replicate each others results if you're not talking about the same thing?
Now I think it's actually a benefit: if your results depend on the minutia of your detector, then is it really a reliable phenomenon?
And the discourse it leads to (why is a result seen in one lab and not another?) leads to better understanding of the phenomena 🙈🐘🙈
@dlevenstein the difficulty I've seen comes when dealing with intracranial data where interictal spikes are really problematic for detecting all of those phenomena, except slow waves. Otherwise I agree as well, would definitely want results to be robust to small differences in detectors.
