Related Experiment Video
Updated: Aug 30, 2025
05:17
Hybrid PET/MRI Imaging of Alzheimer's Disease Based on 18F-AV-1451
Published on: April 18, 2025
328
Mapping the association between tau-PET and Aβ-amyloid-PET using deep learning.
Gihan P Ruwanpathirana1,2, Robert C Williams2, Colin L Masters3,4
1Department of Biomedical Engineering, The University of Melbourne, Melbourne, VIC, Australia.
Scientific Reports
|August 30, 2022
Summary
Alzheimer's disease research reveals how amyloid-beta plaques (Aβ) trigger tau tangles. A novel AI approach using convolutional neural networks (CNNs) identified specific brain regions where Aβ and tau interact, improving our understanding of disease progression.
Area of Science:
- Neuroscience
- Medical Imaging
- Artificial Intelligence in Medicine
Background:
- The molecular mechanisms linking extracellular amyloid-beta (Aβ) deposition to intracellular tau accumulation in Alzheimer's disease (AD) remain unclear.
- Understanding the spatial relationship between Aβ and tau is crucial for elucidating AD pathogenesis.
Purpose of the Study:
- To investigate the association between Aβ burden and tau topography in Alzheimer's disease using advanced neuroimaging and machine learning.
- To compare the efficacy of a convolutional neural network (CNN) approach with traditional linear analysis in modeling Aβ-tau relationships.
Main Methods:
- Utilized a high-resolution Positron Emission Tomography (PET) scanner with low detection thresholds.
- Employed a convolutional neural network (CNN) to analyze the association between Aβ Centiloid values and tau distribution (tau topography).
- Compared CNN findings with standard voxel-wise linear regression analysis.
Main Results:
- The CNN model accurately predicted Aβ Centiloid values from tau topography (R² ranging from 0.72 to 0.86).
- Linear analysis revealed widespread positive correlations between tau and Aβ.
- CNN analysis identified specific focal clusters of Aβ-tau association in the medial temporal lobes, frontal lobes, precuneus, postcentral gyrus, and middle cingulate.
Conclusions:
- The CNN approach identified distinct patterns and regional differences in Aβ-tau associations, varying with Aβ load.
- CNN analysis revealed that middle cingulate, frontal lobe, and precuneus regions are more predictive of Aβ burden at lower Aβ levels, while medial temporal regions are more predictive at higher levels.
- This data-driven CNN method uncovers novel insights into the topographical relationship between tau and Aβ burden in Alzheimer's disease.
