Dear HCP users,
I have some questions about the extensively processed rfMRI data of
“parcellation-timeseries-netmats”. Their filenames are
“groupICA_3T_Q1-Q6related468_MSMsulc.tar.gz”,
“NodeTimeseries_3T_Q1-Q6related468_MSMsulc_ICAd*_ts*.tar.gz”, and
“netmats_3T_Q1-Q6related468_MSMsulc_ICAd*_ts*.tar.gz”. 1. For nodes, is there
any specific threshold applied to the ICA spatial map to obtain the parcel? I
see from an old message of this mailing list that the thumbnail images were
obtained by thresholding the ICA spatial map at Z>6. But how about the
threshold for obtaining that parcellation? Is that also Z>6?
2. These ICA spatial maps were obtained from the ICA+FIX preprocessed data,
therefore, given the dimensionality of 100 as an example, are all these 100
components non- artifactual?
3. For these netmat *.pconn.nii files, are the first column corresponded to the
first ICA component shown in the thumbnail image “0000.png” in the folder
“melodic_IC_sum.sum”, the second column corresponded to the second thumbnail
image “0001.png” in the folder, and so on?
4. For these thumbnail PNG images, could I find any annotation regarding the
corresponding dominant brain region on the HCP website (e.g., something similar
to the annotations in Figure 4(A) of Dr. Smith et al.’s paper
http://www.ncbi.nlm.nih.gov/pubmed/24238796)?
5. Have these data ever been global signal regressed? Many thanks!
_______________________________________________
HCP-Users mailing list
[email protected]
http://lists.humanconnectome.org/mailman/listinfo/hcp-users
