Compute GM/WM contrast on the T2* data to interpret GM CSA values

[jcohenadad]

Based on a suggestion from @renelabounek, the idea would be to compute WM/GM mean signal to understand the origin of the vendor-specific differences in GM CSA, especially from Philips (which shows low values).

To do so, the steps involved are:

• add WM seg (SC minus GM seg)
• extract signal in WM and GM
• output in csv file
• release spine-generic
• re-run processing in data-multi-subject
• compute stats + figs
• create release for data-multi-subject