Analyzing the performance of biomedical time-series
segmentation with electrophysiology data

J 2025

Analyzing the performance of biomedical time-series segmentation with electrophysiology data

REDINA, Richard; Jakub HEJC; Marina FILIPENSKA a Zdeněk STÁREK

Základní údaje

Originální název

Analyzing the performance of biomedical time-series segmentation with electrophysiology data

Autoři

REDINA, Richard; Jakub HEJC; Marina FILIPENSKA a Zdeněk STÁREK

Vydání

Nature Scientific Reports, Berlin, NATURE RESEARCH, 2025, 2045-2322

Další údaje

Jazyk

angličtina

Typ výsledku

Článek v odborném periodiku

Obor

30201 Cardiac and Cardiovascular systems

Stát vydavatele

Německo

Utajení

není předmětem státního či obchodního tajemství

Odkazy

URL

Impakt faktor

Impact factor: 3.900 v roce 2024

Označené pro přenos do RIV

Ano

Kód RIV

RIV/00216224:14110/25:00143872

Organizační jednotka

Lékařská fakulta

Klíčová slova anglicky

Time-series Segmentation; Electrophysiology Study; Rule-based Delineation; Support Vector Machines; U-Net; Faster R-CNN; DENS-ECG

Štítky

14110115, rivok

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 5. 3. 2026 08:20, Mgr. Tereza Miškechová

Anotace

V originále

Accurate segmentation of biomedical time-series, such as intracardiac electrograms, is vital for understanding physiological states and supporting clinical interventions. Traditional rule-based and feature engineering approaches often struggle with complex clinical patterns and noise. Recent deep learning advancements offer solutions, showing various benefits and drawbacks in segmentation tasks. This study evaluates five segmentation algorithms, from traditional rule-based methods to advanced deep learning models, using a unique clinical dataset of intracardiac signals from 100 patients. We compared a rule-based method, a support vector machine (SVM), fully convolutional semantic neural network (UNet), region proposal network (Faster R-CNN), and recurrent neural network for electrocardiographic signals (DENS-ECG). Notably, Faster R-CNN has never been applied to 1D signals segmentation before. Each model underwent Bayesian optimization to minimize hyperparameter bias. Results indicated that deep learning models outperformed traditional methods, with UNet achieving the highest segmentation score of 88.9 % (root mean square errors for onset and offset of 8.43 ms and 7.49 ms), closely followed by DENS-ECG at 87.8 %. Faster R-CNN and SVM showed moderate performance, while the rule-based method had the lowest accuracy (77.7 %). UNet and DENS-ECG excelled in capturing detailed features and handling noise, highlighting their potential for clinical application. Despite greater computational demands, their superior performance and diagnostic potential support further exploration in biomedical time-series analysis.

Návaznosti

MUNI/A/1410/2022, interní kód MU

Název: Nové trendy v diagnostice a managementu srdečních onemocnění

Investor: Masarykova univerzita, Nové trendy v diagnostice a managementu srdečních onemocnění

Citovat

REDINA, Richard; Jakub HEJC; Marina FILIPENSKA a Zdeněk STÁREK. Analyzing the performance of biomedical time-series segmentation with electrophysiology data. Nature Scientific Reports. Berlin: NATURE RESEARCH, 2025, roč. 15, č. 1, s. 1-15. ISSN 2045-2322. Dostupné z: https://doi.org/10.1038/s41598-025-90533-y.

@article{2558418,
   author = {Redina, Richard and Hejc, Jakub and Filipenska, Marina and Stárek, Zdeněk},
   article_location = {Berlin},
   article_number = {1},
   doi = {https://doi.org/10.1038/s41598-025-90533-y},
   keywords = {Time-series Segmentation; Electrophysiology Study; Rule-based Delineation; Support Vector Machines; U-Net; Faster R-CNN; DENS-ECG},
   language = {eng},
   issn = {2045-2322},
   journal = {Nature Scientific Reports},
   title = {Analyzing the performance of biomedical time-series segmentation with electrophysiology data},
   url = {https://www.nature.com/articles/s41598-025-90533-y},
   volume = {15},
   year = {2025}
}

TY  - JOUR
ID  - 2558418
AU  - Redina, Richard - Hejc, Jakub - Filipenska, Marina - Stárek, Zdeněk
PY  - 2025
TI  - Analyzing the performance of biomedical time-series segmentation with electrophysiology data
JF  - Nature Scientific Reports
VL  - 15
IS  - 1
SP  - 1-15
EP  - 1-15
PB  - NATURE RESEARCH
SN  - 20452322
KW  - Time-series Segmentation
KW  - Electrophysiology Study
KW  - Rule-based Delineation
KW  - Support Vector Machines
KW  - U-Net
KW  - Faster R-CNN
KW  - DENS-ECG
UR  - https://www.nature.com/articles/s41598-025-90533-y
N2  - Accurate segmentation of biomedical time-series, such as intracardiac electrograms, is vital for understanding physiological states and supporting clinical interventions. Traditional rule-based and feature engineering approaches often struggle with complex clinical patterns and noise. Recent deep learning advancements offer solutions, showing various benefits and drawbacks in segmentation tasks. This study evaluates five segmentation algorithms, from traditional rule-based methods to advanced deep learning models, using a unique clinical dataset of intracardiac signals from 100 patients. We compared a rule-based method, a support vector machine (SVM), fully convolutional semantic neural network (UNet), region proposal network (Faster R-CNN), and recurrent neural network for electrocardiographic signals (DENS-ECG). Notably, Faster R-CNN has never been applied to 1D signals segmentation before. Each model underwent Bayesian optimization to minimize hyperparameter bias. Results indicated that deep learning models outperformed traditional methods, with UNet achieving the highest segmentation score of 88.9 % (root mean square errors for onset and offset of 8.43 ms and 7.49 ms), closely followed by DENS-ECG at 87.8 %. Faster R-CNN and SVM showed moderate performance, while the rule-based method had the lowest accuracy (77.7 %). UNet and DENS-ECG excelled in capturing detailed features and handling noise, highlighting their potential for clinical application. Despite greater computational demands, their superior performance and diagnostic potential support further exploration in biomedical time-series analysis.
ER  -

REDINA, Richard; Jakub HEJC; Marina FILIPENSKA a Zdeněk STÁREK. Analyzing the performance of biomedical time-series segmentation with electrophysiology data. \textit{Nature Scientific Reports}. Berlin: NATURE RESEARCH, 2025, roč.~15, č.~1, s.~1-15. ISSN~2045-2322. Dostupné z: https://doi.org/10.1038/s41598-025-90533-y.

Přehled o publikaci