Abstract:Individual-level brain parcellation can help accurately depict individual functional organizations, thereby promoting the development of personalized medical practice. In this study, we developed a semi-supervised learning algorithm for the individual-level parcellation based on resting-state fMRI data, and performed the individual precuneus parcellation on three different subjects. The results revealed that the individual-level parcellation exhibited higher connectivity homogeneity compared with the group template and that obvious differences existed in parcellation boundary between individuals, implying the rationality and superiority of the proposed algorithm. This study provided a new approach for individualized brain area parcellation, which may help to better understand the functional characteristics of the human precuneus.

Semi-Supervised Learning for Individual-Level Functional Parcellation of the Human Precuneus