Paolo,

I am traveling right now, so I can only respond briefly (horrible internet connection).

The datasets in the test release will be in the exact same format as the ADHD-200 data already released, with all the same phenotypic variable, minus the diagnostic labels and symptom severity scores (they will have placeholder values [e.g., -99] in the .csv file, so that scripts/code built on the current release format will not be affected).

The test-release data will come from datasets withheld from the initial release (indicated on the site), as well as various datasets currently in collection at the different sites. Data from at least 2 unseen sites will be included.

During the first week of April, more detailed contest specifications will be released, but those are the basics.

Hope that helps.

Best,
Mike