Cerebral Cortex Parcellation via Multiview Clustering of Cortical Thickness Across Lifespan
Poster No:
1645
Submission Type:
Abstract Submission
Authors:
Yinan Feng1, Dan Hu1, Ya Wang1, Jiale Cheng1, Zhengwang Wu1, Li Wang1, Weili Lin1, Gang Li1
Institutions:
1UNC Chapel Hill, Chapel Hill, NC
First Author:
Co-Author(s):
Introduction:
Cortical thickness, a critical marker of brain morphology, undergoes remarkable changes across the human lifespan. While prior studies have explored cortical regionalization during specific life stages (Thambisetty et al., 2010, Wang et al., 2019), the parcellation of the cerebral cortex based on the dynamic development of cortical thickness, spanning from infancy, adolescence, adulthood, and old age, remains poorly understood. To address this gap, we utilized high-resolution MRI data from four large-scale datasets, the Baby Connectome Project (BCP) (Howell et al., 2019), the Lifespan Human Connectome Project Development (HCP-D) (Harms et al., 2018), the Human Connectome Project (HCP) (Van Essen et al., 2013), and the Lifespan Human Connectome Project Aging (HCP-A) (Bookheimer et al., 2019), that cover 0~100 years of age. By treating longitudinal data at each life stage as a distinct view of cortical regionalization and applying advanced multiview clustering techniques, we generated the first fine-grained cerebral cortex parcellation based on cortical thickness spatiotemporal patterns over lifespan, providing novel insights into lifespan brain organization.
Methods:
T1-weighted and T2-weighted structural MR images from 2,294 participants (932 males and 1,362 females, totaling 2,615 scans) from four datasets were included in this study: BCP (283 participants, 604 longitudinal scans, ages 0~6 years), HCP-D (577 participants, ages 6~22 years), HCP (719 participants, ages 22~37 years), and HCP-A (715 participants, ages 36~100 years). Cortical surfaces were reconstructed, aligned, and resampled, followed by the computation of cortical thickness (Wang et al., 2023, Zhao et al., 2021). To leverage shared information across datasets while preserving age-specific characteristics, we employed the Dual Shared-Specific Multiview Subspace Clustering (Zhou et al., 2019), by treating the four datasets as separate views to uncover subspace patterns and group all cortical vertices into k distinct regions, thus avoiding harmonization across datasets. To determine the optimal hyperparameters for multiview clustering, including the dimensions of view-specific and shared features as well as the appropriate number of clusters, we stratified the dataset by age and gender into two partitions and implemented a test-retest framework repeated 50 times. The hyperparameters were evaluated using Silhouette Coefficients (Rousseeuw, 1987), Dice Coefficient, and Adjusted Rand Index (ARI) (Zhou et al., 2019). These three criteria were also utilized to evaluate the final clustering performance, with higher values of these criteria indicating superior parcellations.
Results:
From the evaluation results of the three criteria derived from the clustering illustrated in Fig. 1, local maxima were observed at region numbers 50 and 70. Considering the hemispheric symmetry, spatial distribution, biological interpretability, and the need for a fine-grained yet meaningful representation, we determined 70 regions as the most appropriate configuration for the cerebral cortex parcellation map.
Furthermore, to enhance the anatomical comparability of corresponding regions in both hemispheres, the clustering results were projected to the symmetric fs_LR (32k) surface, facilitating cross-hemisphere comparisons. As depicted in Fig. 2, the resulting parcellation map exhibits bilaterally symmetric patterns and correspond well to regions with known structural and functional significance.
Furthermore, to enhance the anatomical comparability of corresponding regions in both hemispheres, the clustering results were projected to the symmetric fs_LR (32k) surface, facilitating cross-hemisphere comparisons. As depicted in Fig. 2, the resulting parcellation map exhibits bilaterally symmetric patterns and correspond well to regions with known structural and functional significance.
Conclusions:
Leveraging an advanced multiview learning strategy, this study introduces a novel cortical parcellation map that, for the first time, captures the spatiotemporally dynamic development of cortical thickness across the human lifespan. This parcellation offers a unified, consistent, and biologically informed map for comprehensively studying brain development and aging across all stages of life.
Acknowledgement
This work was supported in part by NIH grants (MH123202, ES033518, AG075582, NS128534, and NS135574).
Acknowledgement
This work was supported in part by NIH grants (MH123202, ES033518, AG075582, NS128534, and NS135574).
Lifespan Development:
Early life, Adolescence, Aging 2
Modeling and Analysis Methods:
Segmentation and Parcellation 1
Keywords:
Computational Neuroscience
Cortex
Development
Machine Learning
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Are you Internal Review Board (IRB) certified? Please note: Failure to have IRB, if applicable will lead to automatic rejection of abstract.
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
2. Harms, M. P. (2018). Extending the Human Connectome Project across ages: Imaging protocols for the Lifespan Development and Aging projects. Neuroimage, 183, 972-984.
3. Howell, B. R. (2019). The UNC/UMN Baby Connectome Project (BCP): An overview of the study design and protocol development. NeuroImage, 185, 891-905.
4. Rousseeuw, P. J. (1987). Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics, 20, 53-65.
5. Thambisetty, M.(2010). Longitudinal changes in cortical thickness associated with normal aging. Neuroimage, 52(4), 1215-1223.
6. Van Essen, D. C. (2013). The WU-Minn human connectome project: an overview. Neuroimage, 80, 62-79.
7. Wang, F. (2019). Revealing Developmental Regionalization of Infant Cerebral Cortex Based on Multiple Cortical Properties. International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI), 11765, 841–849.
8. Wang, L. (2023). iBEAT V2. 0: a multisite-applicable, deep learning-based pipeline for infant cerebral cortical surface reconstruction. Nature protocols, 18(5), 1488-1509.
9. Zhao, F. (2021). S3Reg: superfast spherical surface registration based on deep learning. IEEE transactions on medical imaging, 40(8), 1964-1976.
10. Zhou, T. (2019). Dual shared-specific multiview subspace clustering. IEEE transactions on cybernetics, 50(8), 3517-3530.