Skip to main content
. Author manuscript; available in PMC: 2021 Jul 1.
Published in final edited form as: Neuroimage. 2020 Feb 29;214:116678. doi: 10.1016/j.neuroimage.2020.116678

Figure 10-.

Figure 10-

Group Bagging Alone Improves Reproducibility of Within Study/Between Session Subcortical Parcellations. We compare the subcortical parcellations of the 50-minute replication and reference datasets from HNU1 when only group level bagging is performed. The vertical dotted blue line marks the standard clustering approach on the left, and the bagging approach on the right. The horizontal dashed red line shows the highest mean reproducibility of the standard parcellation approach. A & B show the improvement in the spatial correlation of the group level mean adjacency matrix for K = 10 and K = 18, while C & D show the improvement in the ARI of the group cluster labels for K = 10 and K = 18.