To all, This email provides several important updates on the HCP Initial Open Access Data Release announced in mid-October. It is being distributed to the hcp-users mailing list plus the 130 investigators who have agreed to the Data Use Terms.
Corrections and improvements to the processed data (Version 2 forthcoming): If you are using the minimally pre-processed data, please be advised that we have identified an error in the preprocessed BOLD timeseries data: intensity normalization was inadvertently carried out separately on each time frame rather than for the entire time series (i.e., on the 4D volume). In effect, global mean timecourse removal was carried out even though that was not our intent. In addition, we have made several other improvements to the pre-processing pipelines since the October release. This includes improved gradient distortion correction on all modalities, improved EPI distortion correction and better alignment of BOLD volumes to the structural (T1w) volumes. We are currently running data for all 12 subjects through the revised pipelines. We aim to release Version 2 of the unprocessed and processed datasets by late next week, using the same ftp site as before. An updated User Guide will be included with the Version 2 data release. The Version 1 archives on the ftp site will be removed at that time. We apologize for any inconvenience this may have caused. We also stress that the initial data release is intended as an introductory dataset that will let you get familiar with the HCP data quality and format and to become better prepared for the forthcoming Quarter 1 data release next February. Correcting for gradient nonlinearities. If you are using the unprocessed NIFTI datasets, remember that it is important to correct for the spatial distortions caused by gradient nonlinearities associated with the customized Siemens Skyra 3T scanner. These distortions are present in the raw images for ALL modalities (for details, see p. 4 of the User Guide, http://www.humanconnectome.org/documentation/data-release/October2012_Release_User_Guide.pdf). Unfortunately, the gradient field nonlinearity coefficients for the Connectome Skyra required to make these corrections are considered proprietary information by Siemens. This information can be obtained from your Siemens collaboration manager or from Dingxin Wang at [email protected]. It is important that access to these coefficients not be a major obstacle to investigators wanting to make use of the unprocessed HCP datasets. To help us ascertain whether the present arrangement meets the needs of the community, please let us know whether Siemens has been responsive to your requests. FTP downloads. If you have encountered difficulties in getting an HCP account or in downloading archives from the ftp site, please let us know, so that problems can be addressed expeditiously. If the download times differ markedly from what should be expected given the speed of your internet connection, this is also important for us to know. hcp-users. Responses to this email will go to hcp-users, and *not* to the 130 investigators who have agreed to data use terms but have not signed up for hcp-users. If you are in the latter category and wish to subscribe, please do so at https://www.humanconnectome.org/contact/#subscribe. If you are on the hcp-user list but were not previously aware of the HCP Initial data release, you can learn more by visiting the HCP website (http://www.humanconnectome.org/). Thank you for your interest in the HCP and our open access data release. Our consortium has worked extremely hard to generate these data and make them freely available to the neuroimaging and neurscience community. David Van Essen (PI) on behalf of the entire WU-Minn HCP consortium _______________________________________________ HCP-Users mailing list [email protected] http://lists.humanconnectome.org/mailman/listinfo/hcp-users
