Open tsalo opened 3 years ago
We will definitely need to look at the options that other modalities like fMRI are using or proposing, so we could have intermodality agreement. In M/EEG we will work with:
So probably tres
will be the equivalent of time series and freq
of our frequency with no time.
Have they agreed on the data format to use to store this files? I think it is one of our blocking points: see #1
Would a
tres
orfreq
entity be useful for this BEP? I came across an fMRI workflow (rapidtide) where time series are generated with different frequencies, and @effigies pointed out that BEP021 is probably the BEP with the greatest use-case for a corresponding entity.