Related Experiment Video
Updated: Jun 22, 2025

Full- versus Sub-Regional Quantification of Amyloid-Beta Load on Mouse Brain Sections
Published on: May 19, 2022
Generative AI unlocks PET insights: brain amyloid dynamics and quantification.
Matías Nicolás Bossa1, Akshaya Ganesh Nakshathri1, Abel Díaz Berenguer1
1Department of Electronics and Informatics (ETRO), Vrije Universiteit Brussel (VUB), Brussels, Belgium.
Generative AI, using Generative Adversarial Networks (GANs), effectively models brain amyloid accumulation dynamics in Alzheimer's disease (AD) using PET imaging. This approach aids in predicting disease progression and developing new therapies.
Area of Science:
- Neuroimaging
- Artificial Intelligence
- Medical Diagnostics
Background:
- Alzheimer's disease (AD) is characterized by amyloid beta (Aβ) accumulation in the brain.
- Positron Emission Tomography (PET) imaging is vital for visualizing and quantifying Aβ load.
- Understanding the spatiotemporal patterns of Aβ is crucial for AD research and treatment evaluation.
Purpose of the Study:
- To explore the potential of Generative Adversarial Networks (GANs) in modeling brain amyloid dynamics.
- To develop a low-dimensional representation space for brain amyloid load and its temporal changes.
- To demonstrate the utility of generative AI in Alzheimer's disease neuroimaging.
Main Methods:
- Utilized a cohort of 1,259 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) with AV45 PET images.
- Developed a 3D GAN model to project PET images into a latent representation space and generate synthetic images.
- Constructed a progression model using non-parametric ordinary differential equations on the latent space to study Aβ evolution.
Main Results:
- Global Standardized Uptake Value Ratio (SUVR) was accurately predicted from the latent space (RMSE = 0.08 ± 0.01).
- Generated synthetic PET imaging trajectories to simulate Aβ progression over four years.
- Demonstrated the ability to predict and illustrate Aβ changes compared to actual patient progression.
Conclusions:
- Generative AI, specifically GANs, offers powerful tools for statistical prediction and progression modeling in brain amyloid imaging.
- Synthetic patient data and simulated disease trajectories can be generated for research and clinical trial design.
- This study highlights the significant potential of generative AI to advance Alzheimer's disease understanding, diagnosis, and therapeutic development.
More Related Videos
09:31Visualization of Amyloid β Deposits in the Human Brain with Matrix-assisted Laser Desorption/Ionization Imaging Mass Spectrometry
Published on: March 7, 2019
09:33Quantitative 3D In Silico Modeling q3DISM of Cerebral Amyloid-beta Phagocytosis in Rodent Models of Alzheimer's Disease
Published on: December 26, 2016