PhysioEx — Pretrained Sleep Staging Models

Pretrained deep learning models for automatic sleep staging from PhysioEx (Gagliardi et al. 2025).

All models follow the AASM 5-class standard: W, N1, N2, N3, REM.

Available Models

Model	Architecture	Training Dataset	ACC	F1	κ	Params	Reference
sleeptransformer-phan	Transformer	SHHS	0.866	0.796	0.810	2.1M	Phan et al. 2022
tinysleepnet-supratak	CNN + LSTM	Sleep-EDF	0.842	0.790	0.784	83K	Supratak & Guo 2020
seqsleepnet-phan	BiLSTM + Attention	Sleep-EDF	0.810	0.758	0.742	659K	Phan et al. 2019
chambon2018	Braindecode CNN	MASS SS3	0.826	0.731	0.736	29K	Chambon et al. 2018
tsinalis-2016	2-layer CNN	Sleep-EDF	0.773	0.661	0.683	145K	Tsinalis et al. 2016
protosleeptransformer-gagliardi	Prototype-based Transformer	28 sites (multi-dataset)	0.880	0.797	0.829	4.0M	Gagliardi et al.
protosleepnet-gagliardi	Prototype-based BiLSTM + Attention	28 sites (multi-dataset)	0.872	0.780	0.818	792K	Gagliardi et al.

The two *-gagliardi models are trained jointly on 28 sites; ACC / F1 / κ in this table are on the SHHS test subjects (842). Per-site test metrics for every training site are in each model folder (metrics.json) and on the documentation page. These sites were seen in training (other subjects), so the numbers are not an external validation.

Usage

from physioex.models import load_from_pretrained

model = load_from_pretrained("sleeptransformer-phan", verbose=True)

# Extract or download embeddings
from physioex.models import load_embeddings
path = load_embeddings("sleeptransformer-phan", "hmc", verbose=True)

ProtoSleepNet and ProtoSleepTransformer

model = load_from_pretrained("protosleeptransformer-gagliardi", verbose=True)   # L=21; "protosleepnet-gagliardi" has L=20
# input (batch, L, 3, 29, 129): log-power spectrograms of EEG, EOG, EMG; output (batch, L, 5) logits W/N1/N2/N3/REM

Interpretable, prototype-based sleep staging (paper and code, npj Digital Medicine). These checkpoints are the multi-dataset models; they contain no prototype codebook (quantize=True needs one learned after training, see the paper repository). Research use only; not a medical device.

Pre-extracted Embeddings

Model	HuggingFace Repo	Datasets
SeqSleepNet-Phan	4rooms/seqsleepnet-phan-embeddings	25 datasets
TinySleepNet-Supratak	4rooms/tinysleepnet-supratak-embeddings	25 datasets
SleepTransformer-Phan	4rooms/sleeptransformer-phan-embeddings	8 datasets
Chambon2018	4rooms/chambon2018-embeddings	8 datasets
Tsinalis-2016	4rooms/tsinalis-2016-embeddings	8 datasets

Datasets

Dataset	Source	URL
Sleep-EDF	PhysioNet	https://physionet.org/content/sleep-edfx/1.0.0/
HMC	PhysioNet	https://physionet.org/content/hmc-sleep-staging/1.1/
DCSM	ERDA/KU	https://erda.ku.dk/public/archives/db553715ecbe1f3ac66c1dc569826eef/published-archive.html
SHHS	NSRR	https://sleepdata.org/datasets/shhs
MESA	NSRR	https://sleepdata.org/datasets/mesa
HomePAP	NSRR	https://sleepdata.org/datasets/homepap
STAGES	NSRR	https://sleepdata.org/datasets/stages
MASS	CEAMS	http://ceams-carsm.ca/mass/
WSC	NSRR	https://sleepdata.org/datasets/wsc
MrOS	NSRR	https://sleepdata.org/datasets/mros

Citations

@article{gagliardi2025physioex,
    author={Gagliardi, Guido and Alfeo, Luca and Cimino, Mario G C A and Valenza, Gaetano and De Vos, Maarten},
    title={PhysioEx, a new Python library for explainable sleep staging through deep learning},
    journal={Physiological Measurement},
    url={http://iopscience.iop.org/article/10.1088/1361-6579/adaf73},
    year={2025},
}