Yihao Xia

About

I am a Postdoctoral Scholar at the Neuroimaging and Informatics Institute, University of Southern California. I received my Ph.D. in Electrical Engineering from USC (2023), building methods for large-scale diffusion MRI harmonization, groupwise tractography filtering, and related medical image analysis problems.

My current work focuses on medical image harmonization across imaging sources, machine learning for white matter segmentation, and multimodal integration to support reliable, interpretable neuroimaging analyses in collaborative research settings.

I am also interested in statistical methodology for imaging-based studies and in releasing software that makes advanced pipelines easier to use in collaborative research settings.

Diffusion MRI Image harmonization Tractography Brain connectomics Deep learning Multimodal MRI

News

2023.07
Role Started as Postdoctoral Scholar (Research Associate) at the Neuroimaging and Informatics Institute, USC.
2024.03
Paper "Diffusion MRI harmonization via personalized template mapping" published in Human Brain Mapping.
2022.09
Paper "Personalized dMRI harmonization on cortical surface" presented at MICCAI 2022 (Singapore).
2023–
Service Manuscript reviewer for European Journal of Medical Research; IJCARS; IPCAI; ISBI; MICCAI; Med-NeurIPS.

Education

2023
Ph.D. in Electrical Engineering
University of Southern California · Ph.D. conferred June 2023
2017
M.S. in Electrical Engineering
University of Southern California · M.S. awarded May 2017
2013
B.S. in Electrical Engineering
Sichuan University, China · B.S. conferred June 2013

Research

Harmonization & Connectomics

Methods to reduce scanner- and site-related bias in large-scale diffusion MRI cohorts, together with tractography tools that improve bundle consistency across subjects (e.g., groupwise filtering).

Multimodal MRI Integration

Methods that combine complementary imaging signals to improve consistency and interpretability across acquisition protocols and analysis pipelines.

Machine Learning for Segmentation

Deep learning models for structural segmentation (including spine MRI) and software-oriented pipelines that translate methods into usable research tools.

Publications

[J1]
Diffusion MRI harmonization via personalized template mapping.
Xia, Y., Li, Y., & Shi, Y.
Human Brain Mapping, 45(5), e26661, 2024.
[C1]
Personalized dMRI harmonization on cortical surface.
Xia, Y., & Shi, Y.
MICCAI 2022. Lecture Notes in Computer Science, pp. 717–725. Springer, 2022.
[J2]
Groupwise track filtering via iterative message passing and pruning.
Xia, Y., & Shi, Y.
NeuroImage, 221, 117147, 2020.
[J3]
Tractography dissection variability: What happens when 42 groups dissect 14 white matter bundles on the same dataset?
Schilling, K. G., et al. (incl. Xia, Y.)
NeuroImage, 243, 118502, 2021.
[J4]
Multi-parameter ensemble learning for automated vertebral body segmentation in heterogeneously acquired clinical MR images.
Gaonkar, B., Xia, Y., Villaroman, D. S., Ko, A., Attiah, M., Beckett, J. S., & Macyszyn, L.
IEEE Journal of Translational Engineering in Health and Medicine, 5, 1–12, 2017.
[J5]
Surface characterization of ICF capsule by AFM-based profilometer.
Meng, J., Zhao, X., Tang, X., Xia, Y., Ma, X., & Gao, D.
High Power Laser Science and Engineering, 5, e21, 2017.

Contact

I welcome academic collaborations and conversations in diffusion MRI harmonization, brain connectomics, and medical image analysis. Please feel free to reach out.

Neuroimaging and Informatics Institute · 2025 Zonal Ave., Los Angeles, CA 90033