BIDS-iEEG | Where to put reconstructed surface data?

choldgraf commented 6 years ago

We've chatted a few times about the proper place for a surface reconstructions with BIDS-iEEG, and I feel like we should nail this down since @chrisfilo mentioned it in #55 and it'll be a really common component of BIDS-iEEG datasets.

@dorahermes and I were thinking that for iEEG, surfaces are often the key anatomy data we have about a subject (e.g. visualizations are almost never done on volume data any more in ECoG). For that reason we thought they'd go into the main data directory structure, and not in a folder like derivatives/.

However, as @chrisfilo mentions this breaks convention with the rest of the BIDS world. Especially if the goal is to eventually merge the iEEG spec into the core BIDS specification, I think we probably need to put the surfaces in derivatives/ to avoid confusion. @dorahermes what do you think? And does this sound correct @chrisfilo ?

Note: I'm not sure where is the best place to put this question, as we've gone over it a bit in the google doc but I'm worried this isn't a very discoverable place to keep that conversation over time (so feel free to close this issue if people want to nail this down there)

chrisgorgo commented 6 years ago

I had a chat with @dorahermes about this. Here are some thoughts:

I feel that it would be best to keep this consistent with the upcoming Structural preprocessing derivatives extension (see https://docs.google.com/document/d/1YG2g4UkEio4t_STIBOqYOwneLEs1emHIXbGKynx7V0Y/edit#heading=h.nlimrxs2jqsc). This would mean prescribing a path such as /derivatives/<pipeline_name>/sub-<participant_label>/anat/<source_file>_[smoothwm|pial|midthickness|inflated][.L|.R].surf.gii
It's worth noting that since BIDS iEEG specifies the path to the surface file in the JSON file it is not strictly required to prescribe where surface files should be stored. This means that we can postpone this discussion until BIDS Derivatives matures a bit.
We also discussed that the validator should check if either T1w or surface is present. This could be implemented with some conditional logic.

As for where to post such questions. The mailing list is the best place to do this. With so many channels things can get confusing - I tried to make this a bit more clear on the website recently:

choldgraf commented 6 years ago

I'm +1 on punting on this for now, since as Chris mentions it's explicitly required to provide a path to the relevant file anyway. --

On Tue, Feb 27, 2018 at 4:42 PM Chris Filo Gorgolewski < notifications@github.com> wrote:

I had a chat with @dorahermes https://github.com/dorahermes about this. Here are some thoughts:

-

I feel that it would be best to keep this consistent with the upcoming Structural preprocessing derivatives extension (see https://docs.google.com/document/d/1YG2g4UkEio4t_STIBOqYOwneLEs1emHIXbGKynx7V0Y/edit#heading=h.nlimrxs2jqsc). This would mean prescribing a path such as /derivatives//sub-/anat/_[smoothwm|pial|midthickness|inflated][.L|.R].surf.gii

It's worth noting that since BIDS iEEG specifies the path to the surface file in the JSON file it is not strictly required to prescribe where surface files should be stored. This means that we can postpone this discussion until BIDS Derivatives matures a bit.

We also discussed that the validator should check if either T1w or surface is present. This could be implemented with some conditional logic.

As for where to post such questions. The mailing list https://groups.google.com/forum/#!forum/bids-discussion is the best place to do this. With so many channels things can get confusing - I tried to make this a bit more clear on the website recently: [image: image] https://user-images.githubusercontent.com/238759/36763478-01b1c63a-1bdd-11e8-9466-1b55f88f2171.png

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/INCF/BIDS-examples/issues/57#issuecomment-369079367, or mute the thread https://github.com/notifications/unsubscribe-auth/ABwSHdYk8JJbdfeBQT_TipPcLLh9uVf2ks5tZKDYgaJpZM4SVgyX .

dorahermes commented 6 years ago

What does it mean to punt on something - to leave it for now or to make a decision?

choldgraf commented 6 years ago

hehe, sorry for the confusing language :-) I meant "let's hold off on making a decision now, as it sounds like we don't have to make it now, and there will be more information to help us in making it later"

choldgraf commented 6 years ago

currently we're instructing people to put this in derivatives/, so I'm closing this for now unless a new development happens here

bids-standard / bids-examples

BIDS-iEEG | Where to put reconstructed surface data? #57

It's worth noting that since BIDS iEEG specifies the path to the surface file in the JSON file it is not strictly required to prescribe where surface files should be stored. This means that we can postpone this discussion until BIDS Derivatives matures a bit.