This is a collaborative space. In order to contribute, send an email to maximilien.chaumon@icm-institute.org
On any page, type the letter L on your keyboard to add a "Label" to the page, which will make search easier.

Datatype meeting 28.01.2021

Owned by Aurore Bussalb (Unlicensed)
Last updated: Jan 29, 2021 by Maximilien CHAUMON
2 min read

Jan 28, 2021

Attendees

  • Soichi Hayasi

  • @Maximilien CHAUMON

  • Guiomar Niso

  • Brent McPherson

  • @Aurore Bussalb (Unlicensed)

 

  • The datatype can be anything with Brainlife

    • the choice of the datatype is up to the developer

    • a lot of Apps can handle the same datatype

  • It’s possible to deprecate a datatype

    • it is the case for the MEG and EEG datatypes: no too many consequences because few Apps used these datatypes

    • possible to create converters to use the deprecated datatype with an App using the new datatype and vice versa.

    • Datatypes are never deleted. Apps that used a now deprecated datatype will always work (thanks to containers).

  • Generic network datatype is an example of a large datatype that makes up a “standard” for many apps dealing with connectivity analysis to talk a common language. It combines different datatypes (maybe useful for connectivity using MEG and fMRI)

  • Best not to create your own datatype, best to use existing datatypes because Apps will communicate through datatypes.

  • Food for thoughts: In our case, we will think carefully a handful of datatypes (converter datatypes, raw datatype, epochs datatype, source estimation datatype, frequency datatype…)

  • Sometimes an App can be used in a way we don’t expect

  • Datatype is usually linked to a toolbox environment: this shouldn’t be the case in MEG.

  • If in a pipeline we used Apps based on different toolboxes (MNE, Brainstorm, …), converters between Apps may be needed

    • using converters between each App can become heavy

    • in the early steps of processing of a pipeline, converters may not be needed but between more sophisticated steps, converters may be needed

  • No need for an App to read a lot of inputs formats (3 is fine).

  • Brainlife team works on updating how Brainlife inputs MEG and EEG datasets

    • for now, there are 23 datasets EEG and 7 datasets MEG

    • possible to import these datasets in Brainlife

  • Creation of visualization tools for raw MEG and EEG by the Brainlife team

    • Brainstorm visualization tools are very strong

    • but using Brainstorm in Brainlife seems redundant: Guio will start with MNE and she is in contact with the creator of Brainstorm to discuss how Brainstorm Apps can be integrated in Brainlife