A group of US researchers has created a cutting-edge AI tool capable of identifying subtle indicators of Alzheimer’s disease that appear long before a clinical diagnosis is made. These early signs often manifest as irregular behaviors, reflecting initial stages of brain dysfunction.
The team from Gladstone Institutes in California engineered mice to replicate critical aspects of Alzheimer’s and employed a video-based machine learning tool to uncover these early signs of the condition. Their findings, published in Cell Reports, offer a promising new approach for detecting neurological diseases earlier and monitoring their progression over time.
Jorge Palop, a researcher at Gladstone, highlighted that AI could revolutionize how behaviors linked to Alzheimer’s, signaling early brain abnormalities, are analyzed.
The machine learning platform, called VAME (Variational Animal Motion Embedding), analyzed video recordings of mice in an open arena. It detected subtle behavioral changes—such as disorganized actions, unusual patterns, and frequent shifts between different activities—as the mice aged. These behaviors, likely tied to memory and attention deficits, were recorded on camera, though they may not be immediately apparent to the naked eye.
Palop emphasized that this tool could also help track the origins and progression of other neurological diseases. Additionally, the study used VAME to test whether a therapeutic intervention could prevent these abnormal behaviors in Alzheimer’s mice. The team found that blocking a blood-clotting protein called fibrin could prevent the development of these behaviors, even reversing spontaneous changes in Alzheimer’s mice.
