Torsten, Luis,

My take on this (which is probably less well-reasoned than what Ian has already said, but here goes anyway) is that the word "primarily" is important here. My salary was exactly the same after downloading and playing with the challenge data as it was before, and as it would have been had I never downloaded the data. So I would feel fairly confident arguing that my use of the data was not primarily for monetary compensation. If I were employed specifically to segment some brains, then that might be different (probably; I am not a lawyer). If I were to sell original/edited versions of the segmentations for profit, that would clearly be different.

My understanding of the spirit (if not the letter) of the NC clause is that it is typically used by people who want to allow others to use their data, but don't want others to be able to unfairly profit from it (sometimes a particular concern when such exploitation would devalue the content-creator's own activities). As Ian mentioned, the Challenge data was clearly intended to be usable by the MICCAI/imaging community at large, not just the unemployed members of it.

I've tried to find FAQs or other guidance from Creative Commons, but haven't yet found anything directly relevant. Some examples that I have found seem very reasonable, e.g. a school can perform a CC-BY-NC play as long as they don't sell tickets. Presumably the teachers are also paid, but that doesn't mean their use of the script is primarily to get money.

Yours hoping that I (and the many other paid researchers who used the Challenge data) haven't broken the law,
Ged