LongMoE: New Framework for Multimodal Clinical Learning with Missing Data and Patient Trajectories
Researchers have proposed LongMoE, a machine learning framework designed to handle both missing patient data modalities and longitudinal disease progression simultaneously in clinical settings. Existing approaches address these two challenges separately, either ignoring temporal context or assuming complete data availability. The work aims to improve AI-driven clinical decision support for conditions like Alzheimer's disease by making models more robust to real-world data gaps across patient visits over time.
LongMoE (Longitudinal Mixture-of-Experts) is a unified deep learning framework introduced to address two co-occurring problems in multimodal clinical AI: modality missingness, where imaging, text, or health record data may be absent at any given patient visit, and longitudinal dynamics, where the clinical meaning of an observation depends on a patient's disease trajectory over time. Prior methods treat these challenges in isolation, with missing-modality frameworks discarding temporal context and longitudinal models assuming complete data. LongMoE integrates four components: a context-aware imputation module, an attentional tokenization module capturing frequency-domain temporal patterns across irregular visit sequences, a trajectory-aware encoder for modeling disease progression, and a context-conditioned Sparse Mixture-of-Experts router for patient-specific expert selection. The framework was evaluated on three established clinical datasets — ADNI and OASIS-3 (Alzheimer's-focused) and MIMIC-IV (general critical care) — demonstrating improved robustness under missing or weak modalities while remaining competitive when full data is available. The authors present LongMoE as a foundational step toward clinically deployable, longitudinally-aware multimodal learning systems.
What's missing
The preprint has not yet undergone peer review. It is also unclear how LongMoE performs on prospective or out-of-distribution clinical data beyond the three benchmark datasets.
What different sources said
- arXiv cs.LGCenter
LongMoE: Longitudinal Multimodal Learning via Trajectory-Aware Mixture-of-Experts
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.