Datatype meeting 22.04.2021

Date: Apr 22, 2021

Attendees:

• @Aurore Bussalb (Unlicensed)

• Soichi Hayashi

• Franco Pestilli

• Guiomar Niso

• Giulia Berto

• There are several issues with .ctf files, Soichi listed them and their possible solutions in https://docs.google.com/document/d/1WvKbOHU2DuM2CbBp4P141MzBdwyBz0vKqRYDRBneT8Y/edit#. But as the meeting went it turned out that the best solution may be to create conversions Apps:

  • ctf to fif

  • fif to ctf

• Guio and I are going to try to develop it (be careful of the issues highlighted by Soichi regarding .ctf files).

• After running this App, we won’t always keep both files (for instance if we use ctf from OpenNeuro data we will let the ctf files there).

• Note that fif file can’t be bigger than 2GB. If so, it will be split into two files

• Bad channels are written directly in the fif and we don’t want that.

  • we need to create a file with the bad channels

  • put an extra column in channels.tsv or create a specific file with only bad channels in it (the datatype should be updated with it)?

• Where are bad epochs stored?