Mingxia Liu
University of North Carolina at Chapel Hill, NC·Biomedical Engineering
Catching Alzheimer's disease before memory loss becomes obvious would give drugs and interventions their best chance of working. One of the earliest warning signs is subjective cognitive decline (SCD) — when a person notices their own thinking slipping while standard tests still look normal.
- Fields
Biomedical Engineering
Computer Science
Medicine & Clinical Research
- Topics
Alzheimer's disease
subjective cognitive decline
neuroimaging
- Grants
- 1 known grant, through Apr 2027
What they work on
- Develop unsupervised deep learning frameworks that remove scanner- and site-specific differences from brain images, at both the extracted-feature level for precision medicine and the whole-image level for general use, across 79 imaging centers.
- Build a data fusion framework that combines MRI, PET, and cerebrospinal fluid measures without discarding participants who are missing one or more modalities due to dropout or failed scans.
- Create a fast, scalable neuroimage search method that retrieves similar cases from large archives to support subject-specific prediction of progression in subjective cognitive decline and Alzheimer's-related disorders.
- Evaluate all harmonization, fusion, and search tools on a multimodal cohort of more than 5,300 subjects to test whether they reduce bias and improve progression prediction.
More about the research
Studying who among these people will actually progress requires brain scans from thousands of participants, which means pooling data from many hospitals. That pooling creates a technical problem: scanners differ, imaging protocols differ, study populations differ, and different imaging types measure different things. These inconsistencies can bias analyses and lead to conclusions that are artifacts of the equipment rather than the biology. This project builds machine learning tools to clean up and combine such data, and to interpret it for individual patients.
Why it matters
Neuroscience researchers, radiologists, imaging technologists, and study coordinators could use these tools to automatically harmonize and combine brain imaging collected across many hospitals, increasing the statistical power of large studies and reducing conclusions driven by scanner differences rather than biology. Better individual-level prediction of who will progress from subjective cognitive decline toward Alzheimer's disease could support earlier intervention and more efficient drug trials.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Jonathan McConathy
University of Alabama at Birmingham, AL·Neuroscience
Acquire amyloid-PET and tau-PET scans, which detect hallmark Alzheimer's protein deposits, and MRI scans from UAB Alzheimer's Disease Research Center participants using a simultaneous PET/MRI system.
Tracy A. Butler
Weill Medical College of Cornell University, NY·Neuroscience
Confirm, using additional complementary measurement methods, the preliminary finding that traumatic brain injury reduces brain cerebrospinal and interstitial fluid clearance, including the glymphatic pathway, in patients imaged within 14 days of injury.
Spencer L. Bowen
University of Texas Southwestern Medical Center, TX·Biomedical Engineering
Build a device that rapidly and safely moves an external PET source into and out of a PET/MR scanner, so that signal attenuation can be measured directly during patient heart imaging.
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.