Hi Michael, As you proposed, running a modified hcp_fix script that uses the HCP-supplied .fix list of bad components yielded results identical to the denoised data released by the HCP. I'm glad to have such a clear confirmation that my local implementation of ICA-FIX is working correctly! Fully re-running ICA-FIX yielded similar but not identical results, with the same number of estimated components and a maximum difference of only around 700 units, consistent with the expected non-deterministic behavior of MELODIC.
Many thanks for your help, -Ely _______________________________________________ HCP-Users mailing list HCP-Users@humanconnectome.org http://lists.humanconnectome.org/mailman/listinfo/hcp-users
