MGH dataset - How to deal with combined Cord + CSF masks

[MerveKaptan]

Hello everyone,

MGH site kindly shared their 7T fMRI datasets with us.

`

• Their manually drawn masks combine CSF and cord. Again, see Julien's suggestion:

yes, but i suggest we also keep CSF masks because training a 2-class nnUnet to segment the cord and CSF would be very useful. Several people need the CSF mask for fMRI studies. I suggest you or Rohan open an issue to discuss best strategies for these masks (file name, should the CSF mask include SC or should both masks be mutually exclusive, etc)

@rohanbanerjee What do you think?

Thank you!! Merve

[jcohenadad]

@MerveKaptan you described two distinct issues which could (should) be discussed and solved asynchronously. Could you please split this issue into two separate issues. Thx

[MerveKaptan]

Sorry!!! Of course!