Major kudos to this coalition for providing this huge collection of FMRI datasets.
For NIH researchers: We (the SSCC) have downloaded all the datasets that are now available (or was, as of last Friday). We will be making this collection available on the NIH network shortly, as soon as we figure out the mechanics. We are also going to start pre-processing the datasets, to provide registered MNI-ized volumetric time series that can be used for various group analyses.
We have 1061 3.0 Tesla subjects with anatomical and resting state volumes. Data quality varies, and we will be attempting to assess that in the near future.
I am also starting to work on Group InstaCorr, but I think that processing all of this collection interactively will require a pretty nice computer -- I figure it can be done within 32 GB of RAM and with 8+ CPUs -- maybe.