PaperarXiv:2611.08969
PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals
arXiv preprint
TL;DR
Wavelet biosignal FM family (EEG 30M, ECG 30M, sEMG 12M) that folds wavelet subbands before tokenizing (4-7x faster, 7-13x less memory). Top macro-F1 on PTB-XL Super (70.9) and Chapman; reliability-aware fusion lifts SeizeIT2 sensitivity at 11 FA/h from 61.4% to 74.4%. MIT code, Apache-2.0 weights on HF.
Abstract
We present PhysioWave-T, a token-efficient wavelet foundation model family for EEG, ECG, and sEMG. Dynamic scale folding recombines J+1 wavelet subbands prior to tokenization, compressing sequence length from (J+1)CS to CS tokens for J levels, C channels, and S temporal patches. Route-specific front ends and learned channel-name embeddings allow one Transformer per modality to handle diverse montages and sampling rates, while unseen setups reuse the backbone via lightweight adapters. We pre-train a 30M-parameter EEG model on over 15M windows (19–128 channels), alongside ECG (30M) and sEMG (12M) encoders. Across four EEG routes, scale folding cuts Transformer latency by 4.0–7.2x and peak memory by 6.8–12.7x over scale expansion without accuracy loss. Our EEG encoder leads 8 of 9 metrics against BENDR, BIOT, LaBraM, and EEGPT, and surpasses LUNA-L on wearable seizure detection with fewer parameters. Reliability-aware latent fusion (RALF) further integrates summary tokens from independently pre-trained encoders via sample-adaptive weights. On SeizeIT2, RALF raises event sensitivity at 11 false alarms/h from 61.4% to 74.4% and window AUPRC from 0.321 to 0.399 over the best unimodal baseline, outperforming weighted-logit and concatenation fusion across all variants. PhysioWave-T also achieves state-of-the-art sEMG accuracies and top macro-F1 on PTB-XL Super and Chapman-Shaoxing. Code and models: https://github.com/ForeverBlue816/PhysioWave.
Verbatim from arXiv (opens in a new tab) · arXiv metadata (abstracts are released under CC0)
Results on ChatPPG leaderboards · 1
All leaderboards| Leaderboard | Method | Result | Rank |
|---|---|---|---|
| PTB-XL AUROCEcg diagnosis | PhysioWave-T | 90.3AUROC, higher is better | #1 of 1 |
1 row is credited to this paper. Ranks apply within each leaderboard only, as reported by the source.
Tasks & signals
Tasks
Signals
Keywords
Datasets used
Chapman–Shaoxing–Ningbo 12-lead ECG arrhythmia database (PhysioNet)45,152 subjects · ECG- CODE-15%: large annotated 12-lead ECG dataset (Telehealth Network of Minas Gerais)ECG
- CPSC2018 (China Physiological Signal Challenge 2018): 12-lead ECG arrhythmia set, redistributed in PhysioNet Challenge 2020ECG
Icentia11k single-lead continuous raw ECG (PhysioNet)ECG
MedalCare-XL16,900 synthetic 12-lead ECGs from electrophysiological simulations (healthy + 7 pathologies; Zenodo, CC BY 4.0)none (simulated virtual cohort; 16,900 ECGs) · ECG
MIMIC-IV-ECGDiagnostic Electrocardiogram Matched Subset (v1.0)160,000 subjects · ECG
PTB-XL21,799 clinical 12-lead ECGs (PhysioNet)18,869 subjects · ECG
PulseDB v2.0cleaned MIMIC-III + VitalDB cuffless-BP benchmark (MIMIC part ODbL; VitalDB part CC BY-NC-SA)5,361 subjects · PPG · ECG · BP
SeizeIT2wearable behind-the-ear EEG + chest ECG/EMG + accel/gyro (Byteflies Sensor Dot) in 125 focal-epilepsy EMU patients, 886 seizures (OpenNeuro ds005873, CC0)125 subjects · ECG · IMU · EEG
How it works
Cite this paper
Chen, Y., Benatti, S., Benini, L., Guan, C., & Li, Y. (2026). PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals. arXiv preprint. https://arxiv.org/abs/2611.08969
Chen, Yanlong, Simone Benatti, Luca Benini, Cuntai Guan, and Yawei Li. 2026. "PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals." arXiv preprint. https://arxiv.org/abs/2611.08969.
Chen, Y., Benatti, S., Benini, L., Guan, C. and Li, Y. (2026) 'PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals', arXiv preprint. Available at: https://arxiv.org/abs/2611.08969
Chen, Yanlong, et al. "PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals." arXiv preprint, 2026. https://arxiv.org/abs/2611.08969.
Y. Chen, S. Benatti, L. Benini, C. Guan, and Y. Li, "PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals," arXiv preprint, 2026, [Online]. Available: https://arxiv.org/abs/2611.08969.
@misc{chen2026physiowavet,
title = {{PhysioWave-T: Token-Efficient Wavelet Foundation Models for Biosignals}},
author = {Chen, Yanlong and Benatti, Simone and Benini, Luca and Guan, Cuntai and Li, Yawei},
year = {2026},
howpublished = {arXiv preprint},
eprint = {2611.08969},
archivePrefix = {arXiv}
}Added to the ChatPPG research hub