Dec
19
Date: 19-December-2025
Time: 1:00 PM ET (New York Time)
Presenter: Dr. Zhengwu Zhang
Meeting information:
Meeting number: 2539 266 6250
Password: RvCgYhkq387 (78249457 when dialing from a phone or video system)
Join by phone:
+1-415-655-0002 US Toll
Access code: 253 926 66250
https://gsumeetings.webex.com/gsumeetings/j.php?MTID=m0fc0507b0406c29aec987ece0da92c56
Join us December 19th, 2025, 2025, at 1:00 PM ET for an exciting virtual talk by Dr. Zhengwu Zhang entitled: “Continuous and Atlas-free Analysis of Brain Structural Connectivity” as part of the activities of the Brain Space Initiative, co-sponsored by the Center for Translational Research in Neuroimaging and Data Science (TReNDS) and the Data Science Initiative, IEEE Signal Processing Society.
Abstract
Continuous and Atlas-free Analysis of Brain Structural Connectivity
Brain structural networks are commonly summarized as discrete adjacency matrices defined over regions of interest (ROIs) selected from a brain atlas. Because atlas boundaries are coarse and somewhat arbitrary, important sub-ROI connectivity information is often lost. To overcome these limitations, we introduce an atlas-free framework that models structural connectivity as a smooth latent random function defined on a cortical product manifold, referred to as continuous connectivity (ConCon). The high dimensionality of ConCon creates substantial analytical and computational challenges. To address these, we have developed a series of methods that enable efficient ConCon estimation, cross-subject alignment in the continuous domain, and population-level statistical analysis. Together, these tools support fine-grained, localized inference that is not possible with atlas-based connectomes. Using data from the Human Connectome Project and the ABCD Study, we demonstrate that our continuous, atlas-free framework consistently outperforms traditional approaches on multiple structural connectivity analysis tasks and reveals localized cortical patterns associated with meaningful group differences.
Biography
Zhengwu Zhang received the Ph.D. degree in statistics from the Florida State University in 2015 under the supervision of Professor Anuj Srivastav.
He is currently an Associate Professor in Statistics and Operations Research at the University of North Carolina at Chapel Hill and currently serves as an Associate Editor for Reproducibility at the Journal of the American Statistical Association and as an Associate Editor for Biostatistics & Epidemiology. His research develops statistical and machine learning methods for high-dimensional data with low-dimensional structures, with applications spanning finance, artificial intelligence, and neuroscience.
Dr. Zhang’s group works on financial data analysis (predictive modeling, portfolio optimization, LLM applications), LLM reasoning (robust multi-step, causal, and mathematical reasoning), and human cognition and brain research (geometry-based connectome analysis, brain aging, cognitive training, and substance use). His work is supported by multiple NIH grants and reflects a broader mission to design efficient, interpretable, and practical tools for extracting knowledge from complex data.
Recommended Articles:
Wang, L., Li, D., & Zhang, Z. (2025). Riemannian diffusion kernel-smoothed continuous structural connectivity on cortical surface. Imaging Neuroscience, 3, IMAG-a. (Link to Paper).
Cole, M., Xiang, Y., Consagra, W., Srivastava, A., Qiu, X., & Zhang, Z. (2025). Alignment of Continuous Brain Connectivity. arXiv preprint arXiv:2503.15830. (Link to Paper).
Consagra, W., Cole, M., Qiu, X., & Zhang, Z. (2024). Continuous and atlas-free analysis of brain structural connectivity. The Annals of Applied Statistics, 18(3), 1815-1839. (Link to Paper).
Cole, M., Murray, K., St‐Onge, E., Risk, B., Zhong, J., Schifitto, G., ... & Zhang, Z. (2021). Surface‐Based Connectivity Integration: An atlas‐free approach to jointly study functional and structural connectivity. Human Brain Mapping, 42(11), 3481-3499. (Link to Paper).
