AI model reconstructs earlier disease states from later scans

Researchers have developed a novel deep learning model capable of predicting disease progression in reverse, aiming to reconstruct earlier, healthier anatomical states from later diseased scans. This approach addresses a limitation in existing forward-only models, which often miss the crucial incubation period for diseases like Alzheimer's. The proposed two-stage model utilizes a 3D vector-quantised autoencoder and Neural Ordinary Differential Equations to learn continuous-time dynamics, successfully reconstructing unseen previous states on benchmark datasets and outperforming baseline methods on Alzheimer's patient MRIs. AI

IMPACT Enables earlier disease detection and intervention by reconstructing pre-symptomatic states from existing scans.

RANK_REASON The cluster contains an academic paper detailing a new AI model for medical research. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• 3D vector-quantised autoencoder
• Alzheimer's Disease Neuroimaging Initiative
• Morpho-MNIST
• Neural Ordinary Differential Equations
• Ulugbek Shernazarov

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →