MEMP PhD Thesis Defense: Yue Zhi Russ Chua

Foundation models in neuroimaging

Neuroimaging pipelines rely on three core computational tasks—segmentation, deformable registration, and image reconstruction/denoising—to convert MRI scans into quantitative measurements for research and clinical decision-making. Classical methods can be accurate but often require computationally intensive per-scan optimization, while modern deep learning approaches provide fast, amortized inference. However, most deep learning models in neuroimaging are trained for a narrow task and dataset, and performance can degrade under shifts in scanner hardware, acquisition protocols, or label definitions—making it costly to develop and validate new models for each new study. This thesis develops a 3D Neuroimaging Foundation Model (NFM) that is usable out of the box for multiple tasks, while also serving as a reusable starting point for efficient downstream adaptation. The goal is to reduce the effort required to build reliable task-specific tools by learning transferable representations from diverse neuroimaging supervision and enabling practical fine-tuning under limited data and compute. I evaluate NFM on held-out datasets and protocols and compare against strong task-specific baselines and multiple adaptation strategies, with paired statistical testing to support conclusions about performance and data-efficiency.

Thesis Supervisor:
Adrian Dalca, PhD
Director of the Computational Core A.A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital
Assistant Professor of Radiology, Harvard Medical School

Thesis Committee Chair:
John Guttag, PhD
Dugald C. Jackson Professor of Computer Science and Electrical Engineering, MIT

Thesis Readers:
Bruce Fischl, PhD
Professor of Radiology, Harvard Medical School 
Director of the Computational Core A.A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital

William M. Wells III, PhD
Professor of Radiology, Harvard Medical School 
Affiliated Faculty of the Harvard-MIT Division of Health Sciences and Technology
________________________________________________________________________________________

Zoom Invitation
Yue Zhi Russ Chua is inviting you to a scheduled Zoom meeting

Topic: Yue Zhi Russ Chua MEMP PhD Thesis Defense
Time: Wednesday, May 6, 2026, 1:30 PM Eastern Time (US and Canada)

Your participation is important to us: please notify hst [at] mit.edu (hst[at]mit[dot]edu), at least 3 business days in advance, if you require accommodations in order to access this event.

Join Zoom Meeting
https://mit.zoom.us/j/96800246854

One tap mobile
+16465588656,,96800246854# US (New York)
+16699006833,,96800246854# US (San Jose)

Meeting ID: 968 0024 6854

US : +1 646 558 8656 or +1 669 900 6833

International Numbers: https://mit.zoom.us/u/aeG2nKL9gy

Join by SIP
96800246854 [at] zoomcrc.com (96800246854[at]zoomcrc[dot]com)

Join by Skype for Business
https://mit.zoom.us/skype/96800246854