Oscar Esteban

In my opinion, this is more to gauge the possibility that the `SamplingFrequency` metadata is allowed to have a special value (e.g., `"nonregular"`) instead of a `Number` (which is currently...

> the nominal frequency of the recording device I agree with @sjeung and like @effigies' options. The best solution here, in my opinion, is to create an explicit `NominalSamplingFrequency` field....

@sjeung I couldn't find where the timestamps are defined in datetime format. I prefer timestamp because it is generic -- it can be an integer (Unix sense) it can be...

> FWIW I don't like `TabularDataIndex`. I would rather have an inconsistency where physio uses `timestamp` columns and motion uses `LATENCY` channels than make tools have to consult the JSON...

There is an issue with this solution I've just realized. Since some releases ago, I inserted a node at the very beginning of the pipeline to allow datalad datasets. If...

That sounds fine, but as a user, I would prefer a super-simplified report showing a mosaic of the average and a big message saying - this is only three volumes,...

Just a note - something along these lines should also be included in DWI. In the case of DWI, we should not allow less than 6 directions after the removal...

This should be good to go. Merging - thanks Chris.

Please reopen if this is still a problem.

I believe FSL BET just doesn't like these NM T1w images. This would need to continue the roadmap toward an FSL-free implementation.