# More than Alzheimer’s: AI maps the added burden of Lewy body pathology - McKnight Brain Institute

*Источник: McKnight Brain Institute*
*Дата: 2026-07-15*
*Язык: en*

**Кратко:** One challenge to advancing Alzheimer’s disease treatments is that many patients have mixed brain pathologies, meaning more than one disease process is affecting the brain. This can complicate diagnosis, clinical trial design and treatment development because a therapy targeting only one disease process may be less effective if another is also present.

One challenge to advancing Alzheimer’s disease treatments is that many patients have mixed brain pathologies, meaning more than one disease process is affecting the brain. This can complicate diagnosis, clinical trial design and treatment development because a therapy targeting only one disease process may be less effective if another is also present.
Now, University of Florida researchers have used advanced AI with MRI scans and biomarker data from living participants to measure the additional impact of overlapping Alzheimer’s and Lewy body pathologies on brain structure. These findings add to mounting evidence that mixed pathologies are linked to faster neurodegeneration and cognitive decline.
First author Babak Ahmadi, a doctoral student mentored by senior author Abbas Babajani-Feremi, Ph.D., developed, trained and validated a three-dimensional deep-learning model to analyze structural MRI scans and estimate an individual’s MRI-based brain age. Participants were grouped using cerebrospinal fluid measures of Alzheimer’s-related amyloid–tau pathology and an alpha-synuclein seed amplification assay for Lewy body pathology.
“We found that mixed pathology was linked to a substantially heavier MRI-based neurodegenerative burden than either pathology alone,” said Ahmadi. “In other words, when Alzheimer’s and Lewy body pathologies overlap, the brain shows a broader and faster pattern of structural decline.”
Published today in the journal Alzheimer’s & Dementia: The Journal of the Alzheimer’s Association, the results could improve clinical trial design by helping researchers more precisely characterize and group participants for studies of targeted or combination therapies, said Babajani-Feremi, an associate professor in UF’s Department of Neurology, who led the team of researchers from UF’s McKnight Brain Institute and the Norman Fixel Institute for Neurological Diseases at UF Health.
“From post-mortem studies, we know mixed pathologies commonly occur, but being able to identify these disease processes in living patients and measure their structural impact could be key to discovering new integrated interventions to slow cognitive decline, ever more important as our population ages,” Babajani-Feremi said.
Post-mortem studies have shown that about half of autopsy-confirmed Alzheimer’s disease cases also show alpha-synuclein pathology in addition to amyloid and tau, according to the paper.
To investigate whether having mixed Lewy body-Alzheimer’s pathologies accelerates what’s called the “brain-age gap,” or the difference between MRI-predicted brain age and a person’s actual age, Ahmadi developed and validated a deep-learning framework to analyze T1-weighted MRI, a widely used scan for studying brain structure. The model compares MRI-predicted structural age with a person’s actual age, creating a summary measure of neurodegenerative burden.
Drawing from five established large research cohorts, with participant consent, the team first trained the AI model on 4,355 MRI scans from cognitively unimpaired adults to learn typical patterns of structural brain aging. The trained model was then applied to 803 cognitively impaired participants who were grouped using previous cerebrospinal fluid biomarker testing for Alzheimer’s and Lewy body pathologies.
The group with both Alzheimer’s and Lewy body pathology showed the largest deviation in brain-age gap: 6.61 years, compared with the Alzheimer’s-positive/Lewy body-negative group (4.32 years) and the Lewy body-positive/Alzheimer’s-negative group (1.98 years), researchers reported. Over time, the co-pathology group also showed the fastest increase in this structural burden and the steepest tissue loss across several brain regions.
“One of the strengths of this work is that we did not stop at an AI prediction,” Ahmadi said. “We examined which brain regions most influenced the model’s estimates, confirmed structural decline in those regions and connected those changes to cognitive outcomes.”
In addition, the researchers found that females with Alzheimer’s-only pathology or mixed pathology showed higher brain-age gaps than males.
“Previous research suggests that females may be more vulnerable to some Alzheimer’s-related brain changes,” Ahmadi said. “In our work, we saw that this female-specific vulnerability was even more pronounced when Lewy body pathology was also present.”
Moving forward, Babajani-Feremi said researchers are exploring ways to expand the cognitively unimpaired reference cohort to over 50,000 participants to improve the model’s accuracy and generalizability. Another step would be to expand its application from structural MRI to diffusion MRI and functional MRI to capture additional aspects of brain health, with the long-term goal of developing AI-MRI tools to support clinical trial design, risk prediction and treatment planning.al MRI to diffusion MRI and functional MRI to capture additional aspects of brain health, with the long-term goal of developing AI-MRI tools to support clinical trial design, risk prediction and treatment planning.

[Оригинал](https://mbi.ufl.edu/2026/07/15/more-than-alzheimers-ai-maps-the-added-burden-of-lewy-body-pathology/)