Time-Varying EEG Correlations Improve Automated Neonatal Seizure Detection

Tapani, Karoliina; Vanhatalo, Sampsa; Stevenson, Nathan J.

doi:10.1142/s0129065718500302

Cited by 70 publications

(58 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other works have relied on the identification of increasingly complex features to improve system performance. Complex representations such as chaos theory and time-frequency analysis [13], [14], have been explored in an effort to find a single feature which can give separation between seizure and non-seizure classes. A combination of complex features could be more representative of the natural interdependencies in EEG signals.…”

Section: Introductionmentioning

confidence: 99%

Neonatal seizure detection from raw multi-channel EEG using a fully convolutional architecture

et al. 2020

View full text Add to dashboard Cite

A deep learning classifier for detecting seizures in neonates is proposed. This architecture is designed to detect seizure events from raw electroencephalogram (EEG) signals as opposed to the state-of-the-art hand engineered feature-based representation employed in traditional machine learning based solutions. The seizure detection system utilises only convolutional layers in order to process the multichannel time domain signal and is designed to exploit the large amount of weakly labelled data in the training stage. The system performance is assessed on a large database of continuous EEG recordings of 834h in duration; this is further validated on a held-out publicly available dataset and compared with two baseline SVM based systems.The developed system achieves a 56% relative improvement with respect to a feature-based state-of-the art baseline, reaching an AUC of 98.5%; this also compares favourably both in terms of performance and run-time.The effect of varying architectural parameters is thoroughly studied. The performance improvement is achieved through novel architecture design which allows more efficient usage of available training data and end-to-end optimisation from the front-end feature extraction to the back-end classification. The proposed architecture opens new avenues for the application of deep learning to neonatal EEG, where the performance becomes a function of the amount of training data with less dependency on the availability of precise clinical labels.

show abstract

Section: Introductionmentioning

confidence: 99%

Neonatal seizure detection from raw multi-channel EEG using a fully convolutional architecture

et al. 2020

View full text Add to dashboard Cite

show abstract

“…These need to provide high technical quality and well documented multi‐rater classifications for testing human equivalence in diagnostic accuracy, as explained by Tapani et al. . Indeed, the relatively low number of videos with high technical quality was a limitation of our present study and did not allow us to develop a diagnostic classifier of GMs per se .…”

Section: Discussionmentioning

confidence: 90%

Automated pose estimation captures key aspects of General Movements at eight to 17 weeks from conventional videos

Marchi

Hakala²,

Knight³

et al. 2019

Acta Paediatrica

Self Cite

View full text Add to dashboard Cite

show abstract

“…This data descriptor outlines a dataset of multi-channel EEG recordings that have been annotated for neonatal seizures by three independent experts. It was compiled for our recent study on neonatal seizure detection algorithms 19 with parts of the dataset used in another study on inter-observer agreement of neonatal EEG seizure detection 20 . Our dataset can be used as training data for neonatal seizure detection algorithms or as a validation set to compare existing methods.…”

Section: Background and Summarymentioning

confidence: 99%

A dataset of neonatal EEG recordings with seizure annotations

et al. 2019

View full text Add to dashboard Cite

Neonatal seizures are a common emergency in the neonatal intensive care unit (NICU). There are many questions yet to be answered regarding the temporal/spatial characteristics of seizures from different pathologies, response to medication, effects on neurodevelopment and optimal detection. The dataset presented in this descriptor contains EEG recordings from human neonates, the visual interpretation of the EEG by the human experts, supporting clinical data and codes to assist access. Multi-channel EEG was recorded from 79 term neonates admitted to the NICU at the Helsinki University Hospital. The median recording duration was 74 min (IQR: 64 to 96 min). The presence of seizures in the EEGs was annotated independently by three experts. An average of 460 seizures were annotated per expert in the dataset; 39 neonates had seizures and 22 were seizure free, by consensus. The dataset can be used as a reference set of neonatal seizures, in studies of inter-observer agreement and for the development of automated methods of seizure detection and other EEG analyses.

show abstract

Time-Varying EEG Correlations Improve Automated Neonatal Seizure Detection

Cited by 70 publications

References 44 publications

Neonatal seizure detection from raw multi-channel EEG using a fully convolutional architecture

Neonatal seizure detection from raw multi-channel EEG using a fully convolutional architecture

Automated pose estimation captures key aspects of General Movements at eight to 17 weeks from conventional videos

A dataset of neonatal EEG recordings with seizure annotations

Contact Info

Product

Resources

About