Other formats:
BibTeX
LaTeX
RIS
@article{2339682,
author = {Nejedlý, Petr and Kremen, Vaclav and Lepkova, Kamila and Mivalt, Filip and Sladky, Vladimir and Pridalova, Tereza and Plesinger, Filip and Jurak, Pavel and Pail, Martin and Brázdil, Milan and Klimes, Petr and Worrell, Gregory},
article_location = {BERLIN},
article_number = {1},
doi = {http://dx.doi.org/10.1038/s41598-023-27978-6},
keywords = {temporal autoencoder; EEG},
language = {eng},
issn = {2045-2322},
journal = {Nature Scientific Reports},
title = {Utilization of temporal autoencoder for semi-supervised intracranial EEG clustering and classification},
url = {https://www.nature.com/articles/s41598-023-27978-6},
volume = {13},
year = {2023}
}
TY - JOUR
ID - 2339682
AU - Nejedlý, Petr - Kremen, Vaclav - Lepkova, Kamila - Mivalt, Filip - Sladky, Vladimir - Pridalova, Tereza - Plesinger, Filip - Jurak, Pavel - Pail, Martin - Brázdil, Milan - Klimes, Petr - Worrell, Gregory
PY - 2023
TI - Utilization of temporal autoencoder for semi-supervised intracranial EEG clustering and classification
JF - Nature Scientific Reports
VL - 13
IS - 1
SP - 1-13
EP - 1-13
PB - NATURE
SN - 20452322
KW - temporal autoencoder
KW - EEG
UR - https://www.nature.com/articles/s41598-023-27978-6
N2 - Manual visual review, annotation and categorization of electroencephalography (EEG) is a time-consuming task that is often associated with human bias and requires trained electrophysiology experts with specific domain knowledge. This challenge is now compounded by development of measurement technologies and devices allowing large-scale heterogeneous, multi-channel recordings spanning multiple brain regions over days, weeks. Currently, supervised deep-learning techniques were shown to be an effective tool for analyzing big data sets, including EEG. However, the most significant caveat in training the supervised deep-learning models in a clinical research setting is the lack of adequate gold-standard annotations created by electrophysiology experts. Here, we propose a semi-supervised machine learning technique that utilizes deep-learning methods with a minimal amount of gold-standard labels. The method utilizes a temporal autoencoder for dimensionality reduction and a small number of the expert-provided gold-standard labels used for kernel density estimating (KDE) maps. We used data from electrophysiological intracranial EEG (iEEG) recordings acquired in two hospitals with different recording systems across 39 patients to validate the method. The method achieved iEEG classification (Pathologic vs. Normal vs. Artifacts) results with an area under the receiver operating characteristic (AUROC) scores of 0.862 +/- 0.037, 0.879 +/- 0.042, and area under the precision-recall curve (AUPRC) scores of 0.740 +/- 0.740, 0.714 +/- 0.042. This demonstrates that semi-supervised methods can provide acceptable results while requiring only 100 gold-standard data samples in each classification category. Subsequently, we deployed the technique to 12 novel patients in a pseudo-prospective framework for detecting Interictal epileptiform discharges (IEDs). We show that the proposed temporal autoencoder was able to generalize to novel patients while achieving AUROC of 0.877 +/- 0.067 and AUPRC of 0.705 +/- 0.154.
ER -
NEJEDLÝ, Petr, Vaclav KREMEN, Kamila LEPKOVA, Filip MIVALT, Vladimir SLADKY, Tereza PRIDALOVA, Filip PLESINGER, Pavel JURAK, Martin PAIL, Milan BRÁZDIL, Petr KLIMES and Gregory WORRELL. Utilization of temporal autoencoder for semi-supervised intracranial EEG clustering and classification. \textit{Nature Scientific Reports}. BERLIN: NATURE, 2023, vol.~13, No~1, p.~1-13. ISSN~2045-2322. Available from: https://dx.doi.org/10.1038/s41598-023-27978-6.
|
