MINT: Multimodal Imaging-to-Speech Knowledge Transfer for Early Alzheimer's Screening

Researchers developed MINT, a framework that transfers knowledge from MRI brain scans to speech analysis for early Alzheimer's detection. The system achieves comparable performance to speech-only methods while being grounded in neuroimaging biomarkers, enabling population-scale screening without requiring expensive MRI scans at inference.

• →MINT framework enables early Alzheimer's screening through speech analysis without requiring MRI scans during actual testing.
• →The system trains on 1,228 subjects and achieves AUC of 0.720 for detecting mild cognitive impairment versus normal cognition.
• →Multimodal fusion combining MRI and speech data improves performance to 0.973 AUC compared to 0.958 for MRI alone.
• →This represents the first demonstration of MRI-to-speech knowledge transfer for Alzheimer's screening applications.
• →The approach could enable population-scale cognitive screening by eliminating the need for expensive neuroimaging infrastructure.