Original title: ProMUSE: Progressive Multi-modal Uncertainty-guided Staged Evidential Alzheimer Disease Classification
Long Doan, Branden Chen, Ethan Litton, H Huang +5 · October 2026
Diagnosing Alzheimer's disease often uses cheap clinical tests along with expensive brain scans. This new system starts by checking only the low-cost clinical records. It measures how confident it is in that first guess. If the system is too uncertain, it asks for a magnetic brain scan or another specialized scan. By combining the separate clues only when necessary, it keeps accuracy high while avoiding costly scans for most patients.
You are packing a backpack for a day out, trying to keep it as light as possible. Shoving in heavy rain gear every single morning is exhausting, but leaving it behind means you might get soaked. Here, the heavy gear corresponds to expensive brain scans, and the light jacket is the standard clinical paperwork. The system checks the basic forecast first and only packs the heavy gear when the risk of being wrong is too high.
Treating Alzheimer's disease works best early, but confirming it usually requires expensive brain scans that many clinics cannot afford. This system helps doctors rely on simpler clinical assessments first, ordering costly scans only when the diagnosis is uncertain. That makes early screening accessible to far more patients at a fraction of the cost.
The abstract does not report testing outside the ADNI, AIBL, and OASIS datasets, nor does it evaluate performance across conditions other than the specified CN-AD, CN-MCI, and MCI-AD classification tasks.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
ProMUSE: Progressive Multi-modal Uncertainty-guided Staged Evidential Alzheimer Disease Classification — Long Doan, Branden Chen, Ethan Litton et al. — Intelligent Computing — doi:10.34133/icomputing.1555
Abstract licence: CC BY
Explanation generated by Synthesis from the abstract — it is not the original paper.