xqi10@hawk.iit.edu

The purpose of this work was two-fold: a) to enhance the white matter connectome of the IIT Human Brain Atlas through an improved tractography strategy and use of more precisely defined gray matter labels and b) to evaluate how representative the new connectome is of young adult participants of the Human Connectome Project (HCP).

Data: i) the HARDI and DTI templates of the IIT Human Brain Atlas (v.4.1), ii) more precisely defined gray matter labels generated for the atlas and iii) the structural and diffusion MRI preprocessed data of 20 unrelated young adult HCP participants with balanced sex and age. Connectome construction: Whole brain anatomically-constrained tractography, spherical-deconvolution informed filtering (SIFT), and 84 Desikan-Killiany regions were applied on the IIT HARDI template and 20 HCP datasets to generate 21 connectomes, each containing 6972 edges. Connectome evaluation: The streamlines of the edges survived filtering of each connectome were transformed to the other 20 connectome spaces by DTI registration and the tract density images (TDI) were generated. F1 scores for each edge's spatial volumes between one connectome and the other 20 connectomes were computed. Pearson's correlation coefficients were also computed for TDI maps between all 20 pairs. These processes were repeated for each connectome in their spaces and the F1 scores and the Pearson's correlations were averaged. One sample t-tests were used to test the IIT connectome if the average F1 score and Pearson's correlation were significantly different than those of the HCP participants.

TDI and edge density images of the new IIT connectome has shown good correspondence with other studies. The IIT connectome has no significantly different average F1 scores and significantly higher average Pearson's correlations (p<0.0001) than those of the HCP group over the different numbers of survived edge, indicating that the IIT connectome is representative of individual HCP connectomes.

A new version of the IIT white matter connectome was developed and compared to those of 20 young adult HCP participants. It was demonstrated that the IIT connectome is representative of individual HCP connectomes.

The new IIT white matter connectome will facilitate atlas-based analyses of the whole brain connectome.