New Foundation Model 'Hypnos' Uses Next-Token Prediction to Learn Sleep Physiology Representations
Researchers have developed Hypnos, a multi-modal sleep foundation model trained on over 20,000 overnight polysomnography recordings using a next-token prediction approach. The model processes eight sensing modalities—including EEG, ECG, and respiratory signals—and outperforms existing foundation models on sleep-related benchmarks. It achieves supervised-level accuracy in sleep stage classification using 100 times less labeled data and also generalizes to detecting atrial fibrillation from daytime ECG data.
Hypnos is a large auto-regressive foundation model designed to learn generalizable representations from multi-modal physiological signals, with a primary focus on sleep medicine. Unlike prior approaches that relied on masked-reconstruction or contrastive learning objectives, Hypnos uses next-token prediction—the same core technique underlying large language models—applied to discrete token streams derived from eight sensing modalities via residual vector quantization. The model was trained on more than 20,000 overnight polysomnography recordings and can operate on any supported subset of modalities at inference time, making it flexible for real-world clinical settings. On sleep stage classification benchmarks, Hypnos matches strong supervised baselines while requiring only one percent of the labeled data those baselines need. Notably, the model also generalizes beyond sleep physiology, outperforming a dedicated ECG foundation model on atrial fibrillation detection from daytime recordings. The authors argue these results establish next-token prediction as a strong self-supervised objective for physiological signal representation learning, with potential applications extending to cardiology, neurology, and other healthcare domains.
What's missing
The paper does not report external clinical validation on prospective or real-world deployment datasets beyond held-out test splits from the training distribution. Key open questions include how Hypnos performs across demographic subgroups (age, sex, comorbidities), whether the 20,000-recording training set introduces dataset-specific biases, and how the model behaves with lower-quality or consumer-grade sensor inputs rather than clinical polysomnography equipment. The computational cost of training and inference relative to supervised baselines is not detailed in the abstract.
What different sources said
- arXiv cs.AICenter
Next-Token Prediction Learns Generalisable Representations of Sleep Physiology
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.