Photo: Pat Sargent
A recent study by the Laboratory for Neuro-Analysis and Imaging (LANAI) at UMass Chan Medical School highlights how subcortical brain structures degenerate in Alzheimer’s disease and frontotemporal dementia patients.
The study, published in the Journal of Alzheimer’s Disease, is led by Manojkumar Saranathan, PhD, professor of radiology, head of the LANAI and adjunct professor of biomedical engineering at UMass Amherst; and Arit Banerjee, a research assistant in Dr. Saranathan’s lab and a PhD candidate in bioengineering and biomedical engineering at UMass Amherst.
Over the past several years, Saranathan has developed a brain mapping tool he calls subcortical-Thalamus-Optimized Multi-Atlas Segmentation (sTHOMAS). The tool can resolve exceptionally fine subcortical brain regions using standard MRI scans. In September 2025, Saranathan published the sTHOMAS software available for download in Human Brain Mapping, along with findings of segmentation of thalamic nuclei from structural MRI.
This study used sTHOMAS to examine deep gray matter brain structures that are responsible for sensory and motor control and how these structures degenerate differently in Alzheimer’s disease (AD) compared to frontotemporal dementia (FTD), and how these changes relate to cognition and diagnosis. They analyzed MRI scans from 380 Alzheimer’s Disease Neuroimaging Initiative participants and 274 Frontotemporal Lobar Degeneration Neuroimaging Initiative participants, with longitudinal follow-up in 237 individuals.
“We’ve been developing methods for visualization and characterization of these deeper brain structures because they haven’t been easily visible. Now that the tools are in place, we were interested in looking at Alzheimer’s disease and FTD. What we were interested in is characterizing these structures and how they decay in these two different diseases, and see if there are different patterns,” Saranathan said.
They found that Alzheimer’s disease and FTD scans both showed degeneration very early in areas of the thalamus, nucleus accumbens and claustrum using cross-sectional, longitudinal and machine-learning analyses. However, the atrophy was stronger and the decline in deep gray matter was more rapid in FTD patients.
Their findings may ultimately provide ways to identify different forms of dementia and track how quickly they progress in subcortical regions. Saranathan and Banerjee aim to identify precisely when the subcortical brain structures are affected by these forms of dementia.
“What’s exciting in our research is that while this study focuses on classification, it lays the groundwork for true future prediction,” Banerjee said. “It’s possible in the future a patient may walk into a clinic and we could already anticipate, years before symptoms appear, whether they are likely to develop Alzheimer’s or FTD.”
Banerjee, the first author on the study, presented the study’s findings at the Department of Neurology’s Neurology Nexus meeting series in November. Co-authors on the study include Joyita Dutta, PhD, professor of biomedical engineering at UMass Amherst; Fan Yang, PhD, postdoctoral fellow in biomedical engineering at UMass Amherst; Alberto Cacciola, associate professor of human anatomy at Humanitas University in Milan; and Michael Hornberger, professor of applied dementia research at the University of Southampton.
