Deep Multi-Modal Detection of Early Alzheimer’s Disease (2025-2026)
Background
According to the Alzheimer’s Association, about 6.9 million people in the U.S. have Alzheimer’s disease (AD). That number is projected to jump to 12.7 million by 2050 as the elderly population grows. AD is among the leading causes of death for older Americans and disproportionately impacts women and Black and Hispanic individuals.
Brain changes associated with AD can begin developing many years before noticeable symptoms appear. Early and accurate diagnosis has become increasingly critical, particularly with the introduction of new treatments for AD. In this evolving landscape, MRI-based biomarkers have emerged as a key tool for detecting the disease at its early stages, assessing the effectiveness of therapies and predicting treatment outcomes. Identifying these biomarkers is essential for enabling timely intervention and improving the overall management of AD.
Project Description
This project team will identify novel MRI-based biomarkers that can improve early Alzheimer’s disease diagnosis and predict disease progression by detecting subtle patterns and alterations in brain topology. The team will apply methods from computational topology and geometry to extract shapes and features from diffusion MRI (dMRI) and functional MRI (fMRI) data, utilizing large datasets from the Alzheimer’s Disease Neuroimaging Initiative and the National Alzheimer’s Coordinating Center.
By integrating dMRI and fMRI data, the team aims to capture complementary aspects of brain network organization. They will then correlate and analyze these topological features with key AD biomarkers, such as AD pathology, cognitive performance and disease stages, using robust statistical techniques and machine learning to assess their biological underpinnings and interpretations. The project’s ultimate goal is to develop predictive frameworks that handle high-dimensional, multi-source data while accounting for variables like age and sex.
Anticipated Outputs
Open-source codebase for the MRI-based biomarker framework; peer-reviewed publications; conference presentations; data to guide future research
Student Opportunities
Ideally, this project team will include 2 graduate students and 4 undergraduate students interested in computer science, mathematics, neuroscience, biostatistics/bioinformatics and/or engineering. The strongest applicants would have some experience in quantitative methods and programing; familiarity with Python will be considered an added benefit. Two graduate students with excellent management skills and preferably technical expertise will be selected to serve as project managers.
Students will conduct literature reviews, analyze data related to Alzheimer’s biomarkers using data analytics tools and engage in team-based problem solving. Graduate students will gain valuable mentorship experience. Ultimately, all team members will develop expertise in research methods, data analysis, teamwork and communication.
See the related Data+ project for Summer 2025; there is a separate application process for students who are interested in this optional component.
Timing
Summer 2025 – Spring 2026
- Summer 2025 (optional): Begin developing deep learning model; extract topological features from preprocessed fMRI data and complement these with established clinical biomarkers
- Fall 2025: Explore the biological underpinnings of topological patterns in functional connectomes; obtain preliminary results; engage in visualization, interpretation and presentation of findings
- Spring 2026: Integrate structural data into functional analyses; compare results to single-modality analyses; compile overall research outcomes into comprehensive documentation; present findings; possibly draft peer-reviewed article
Crediting
Academic credit available for fall and spring semesters; summer funding available
See related Data+ summer project, Deep Multimodal Detection of Early Alzheimer’s Disease (2025), and earlier related team, Analyzing Alzheimer's Biomarkers Through Dynamic Brain Topology (2024-2025).