ICA parcellation from HCP

[neo]

HCP S1200 provides brain parcellations from group ICA with different resolutions (15, 25, 50, … clusters). However from its visualization, it is so broken (one hemisphere below) and far from what I expected. Is it really a brain parcellation?

file path: HCP1200_Parcellation_Timeseries_Netmats.zip → groupICA → groupICA_3T_HCP1200_MSMAII_d15.ica → molodic_IC_ftb.dlabel.nii

[JoeBathelt]

Just in case anyone stumbles over this post on Google, it looks like the parcellation is not plotted on the correct surface. The HCP data is on the HCP average surface and the plot seems to be on the fsLR surface. When plotted on that surface, the ICA parcellation looks fine: