Abstract
Alzheimer’s disease (AD) is a multifactorial neurodegenerative disorder arising from complex interactions among genetic susceptibility, environmental exposures, and brain pathology. Yet most computational approaches to AD diagnosis rely primarily on neuroimaging-derived phenotypes, overlooking the upstream genetic and environmental factors that shape disease onset and progression. Moreover, these methods typically train a single shared model across all individuals, implicitly assuming a common disease mechanism. This assumption is at odds with the marked heterogeneity of AD, in which similar clinical symptoms can arise from distinct genetic, environmental, and neuropathological pathways—variability that a single shared model struggles to capture. To address this, we propose SAGE-MoE, a Subject-Adaptive Genetic-Environmental guided Mixture-of-Experts framework for early diagnosis of AD. SAGE-MoE integrates multimodal (structural and functional) brain connectivity with genetic and environmental signals that dynamically route each subject to a sparse subset of specialized experts. By modeling heterogeneous disease mechanisms through subject-conditioned routing, SAGE-MoE learns individualized diagnostic representations and improves normal control (NC) versus mild cognitive impairment (MCI) classification on the ADNI cohort.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/MLMI_043.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=gH5vq8U5rS
BibTex
@InProceedings{GuJin_SAGEMoE_MICCAISAT2026,
author = { Gu, Jinchen AND Lyu, Yanjun AND Zhao, Nan AND Qiu, Lei AND Zhang, Lu},
title = { { SAGE-MoE: Subject-Adaptive Genetic-Environmental Guided Mixture-of-Experts for Early Diagnosis of Alzheimer’s Disease } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17272},
month = {pending},
page = {pending}
}