Quefeng Li
University of North Carolina at Chapel Hill, NC·Mathematics & Statistics
Brain studies of Alzheimer's disease (AD) now collect huge amounts of imaging data. These data often include several types of brain scans for each person (multimodal data), repeated over time and pooled from several study groups (multi-cohort data).
- Fields
Mathematics & Statistics
Public Health & Epidemiology
Neuroscience
- Topics
statistical methods
biostatistics
integrative analysis
- Grants
- 1 known grant, through Jun 2027
What they work on
- Develop a factor-adjusted integrative model for multimodal neuroimaging data within an M-estimation framework, with inference tools to test the importance of a whole imaging type or of combined predictors across types.
- Create a computational tool that handles block-missing multimodal data without guessing missing values, using a mini-batch gradient descent algorithm to produce reliable estimates.
- Build an interactive factor model for longitudinal data from multiple cohorts that includes the standard random effects model as a special case and better accounts for differences between people.
More about the research
Combining these data could reveal image-based markers for diagnosing AD early. But rigorous statistical methods and efficient computing tools for this combined, or integrative, analysis are lacking. Two problems remain unsolved: whole blocks of data are often missing, and people differ from one another in ways models must account for.
Why it matters
Alzheimer's disease researchers increasingly combine brain scans from many people, imaging types and studies, but missing data and differences between people limit what they can conclude. These statistical tools could help find imaging markers for early AD diagnosis and improve understanding of disease mechanisms. The methods could also be applied to other biomedical studies that combine data sources.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Yao Zheng
University of Connecticut, CT·Mathematics & Statistics·Early-career
Develop recurrent neural network (RNN) models for high-dimensional time series that stay interpretable through Granger causality, a test of whether one series predicts another, to balance accuracy and interpretability.
Benjamin Boniece
Drexel University, PA·Mathematics & Statistics
Develop methods to detect shocks that affect many series in high-dimensional, high-frequency data, possibly asynchronously, using time-localized principal component analysis in factor-structured models.
Hok Chio Lai
University of Southern California, CA·Psychology·Early-career
Develop a robust optimization algorithm for data harmonization that puts different instruments measuring the same concept on a common scale while also correcting for measurement bias across groups.
The school
University of North Carolina at Chapel Hill: 72 labs on PI Pages, $1.6B a year on research, 523 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.