# Uncovering Trajectory and Topological Signatures in Multimodal Pediatric Sleep Embeddings

arXiv: https://arxiv.org/abs/2605.14156 · HTML: https://mlsanity.com/paper/2605.14156/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 35 references that are in the map.
- First posted on arXiv 13 May 2026.

## Builds on (most cited first, 9 of 9)

- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Persistence Images: A Stable Vector Representation of Persistent Homology](https://mlsanity.com/paper/1507.06217/) (2015)
- [DeepSleepNet: a Model for Automatic Sleep Stage Scoring based on Raw Single-Channel EEG](https://mlsanity.com/paper/1703.04046/) (2017)
- [Uncovering the structure of clinical EEG signals with self-supervised learning](https://mlsanity.com/paper/2007.16104/) (2020)
- [XSleepNet: Multi-View Sequential Model for Automatic Sleep Staging](https://mlsanity.com/paper/2007.05492/) (2020)
- [SleepFM: Multi-modal Representation Learning for Sleep Across Brain Activity, ECG and Respiratory Signals](https://mlsanity.com/paper/2405.17766/) (2024)
- [A Large Collection of Real-world Pediatric Sleep Studies](https://mlsanity.com/paper/2102.13284/) (2021)
- [Automatic Sleep Scoring from Large-scale Multi-channel Pediatric EEG](https://mlsanity.com/paper/2207.06921/) (2022)
- [PedSleepMAE: Generative Model for Multimodal Pediatric Sleep Signals](https://mlsanity.com/paper/2411.00718/) (2024)

## Abstract (arXiv)

While generative models have shown promise in pediatric sleep analysis, the latent structure of their multimodal embeddings remains poorly understood. This work investigates session-wide diagnostic information contained in the sequences of 30-second pediatric PSG epochs embedded by a multimodal masked autoencoder. We test whether augmenting embeddings with PHATE-derived per-epoch coordinates and whole-night movement descriptors, persistent homology summaries of the embedding cloud, and EHR yields task-relevant signals. Simple linear and MLP models, chosen for interpretability rather than state-of-the-art performance, show that geometric, topological, and clinical features each provide complementary gains. For binary predictions, feature importance is task-dependent, and more expressive late-fusion models generally perform better, with AUPRC improving from 0.26 to 0.34 for desaturation, 0.31 to 0.48 for EEG arousal, 0.09 to 0.22 for hypopnea, and 0.05 to 0.14 for apnea. We also report Brier score and Expected Calibration Error, where the full fusion model yields the best calibration across all four binary tasks. Our study reveals that latent geometry/topology and EHR offer complementary, interpretable signals beyond embeddings, improving calibration and robustness under extreme imbalance.