About
I am a postdoctoral researcher in the NISOx group at the University of Oxford, working with Prof. Thomas Nichols. I previously received my Ph.D. in Health Data Science from the same group, supported by the National Institutes of Health (NIH) studentship, during which I was also affiliated with the Department of Computer Science and the Nuffield Department of Population Health.
My research focuses on statistical methods for neuroimaging, including meta-analysis of fMRI data, spatial modelling of brain lesions, and applications of machine learning in neuroscience. A central theme of my work is developing new statistical frameworks for neuroimaging data analysis, with a particular emphasis on coordinate-based meta-analysis and spatial modelling approaches.
Education
Ph.D. in Health Data Science
🎓 University of Oxford
- • Supervised by Prof. Thomas Nichols
- • NIH studentship funded
- • Affiliated with Computer Science & Population Health
BSc Mathematics (First Class)
🎓 University of Manchester
- • Ranked #2/340
- • Dissertation advised by Prof. Korbinian Strimmer and Prof. Francoise Tisseur
- • Maths Excellence Awards
- • International Student Scholarship
Research Projects & Publications
📊 Efficient Lesion Estimation Using a Spatial Poisson Process and a Scalable Approximate Model
Yifan Yu, Thomas Nichols
We propose a novel approach to investigate the association between lesion incidence and spatial varying risk factors, such as age and CVR factors, and developed a robust, computationally efficient and scalable statistical inference framework.
🧠 Meta Regression and Inference for fMRI Data Across Multiple Groups
Yifan Yu, Lauren Hill-Bowen, Michael Cody Riedel, Angela Laird, Thomas Nichols
Under reviewWe extend Coordinate-based meta-regression (CBMR) to allow for different stochastic models and build a statistical inference framework using bootstrapping approach. This new approach allows both group-specific and group comparison of brain activation intensity estimations.
📈 Coordinate-Based Meta-regression of fMRI data
Yifan Yu, Rosario Pintos Lobo, Michael Cody Riedel, Angela Laird, Thomas Nichols
Published on BiostatsWe develop a novel Meta-regression framework, as a computationally efficient alternative to previous Bayesian spatial regression model. We propose different variants of voxel-wise or study-wise statistical distributions to find the most accurate but parsimonious model.
🏆 Convergence Properties and Sensitivity of the PageRank Problem
Yifan Yu, Françoise Tisseur
Best Dissertation AwardWe compare the computational efficiency and stability of several numerical methods in the PageRank system, and we also investigate the sensitivity of teleportation parameter with perturbation method.
Work Experience
🔬 Research Scientist Intern
Amazon
📍 Luxembourg, Luxembourg
Worked on optimization models for Amazon EU transportation network, conducting backtesting and improving computational efficiency through stochastic modeling approaches.
📊 Data Analyst Intern
Hithink RoyalFlush
📍 Hangzhou, China
Administered user databases with SQL, analyzed user activity strategies, and developed web crawlers for recommendation system improvements.
Contact
Get in Touch
Location
🏛️ University of Oxford
Big Data Institute
Old Road Campus
Oxford, UK