Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG
Abstract Neonatal seizures require urgent treatment, but often go undetected without expert EEG monitoring. We have developed and validated a seizure detection model using retrospective EEG data from 332 neonates. A convolutional neural network was trained and tested on over 50,000 hours (n = 202) o...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | npj Digital Medicine |
| Online Access: | https://doi.org/10.1038/s41746-024-01416-x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841544328921481216 |
|---|---|
| author | Robert Hogan Sean R. Mathieson Aurel Luca Soraia Ventura Sean Griffin Geraldine B. Boylan John M. O’Toole |
| author_facet | Robert Hogan Sean R. Mathieson Aurel Luca Soraia Ventura Sean Griffin Geraldine B. Boylan John M. O’Toole |
| author_sort | Robert Hogan |
| collection | DOAJ |
| description | Abstract Neonatal seizures require urgent treatment, but often go undetected without expert EEG monitoring. We have developed and validated a seizure detection model using retrospective EEG data from 332 neonates. A convolutional neural network was trained and tested on over 50,000 hours (n = 202) of annotated single-channel EEG containing 12,402 seizure events. This model was then validated on two independent multi-reviewer datasets (n = 51 and n = 79). Increasing data and model size improved performance: Matthews correlation coefficient (MCC) and Pearson’s correlation (r) increased by up to 50% (15%) with data (model) scaling. The largest model (21m parameters) achieved state-of-the-art on an open-access dataset (MCC = 0.764, r = 0.824, and AUC = 0.982). This model also attained expert-level performance on both validation sets, a first in this field, with no significant difference in inter-rater agreement when the model replaces an expert (∣Δ κ∣ < 0.094, p > 0.05). |
| format | Article |
| id | doaj-art-f30baebb107e435180eb462269de4f9f |
| institution | Kabale University |
| issn | 2398-6352 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | npj Digital Medicine |
| spelling | doaj-art-f30baebb107e435180eb462269de4f9f2025-01-12T12:40:55ZengNature Portfolionpj Digital Medicine2398-63522025-01-018111110.1038/s41746-024-01416-xScaling convolutional neural networks achieves expert level seizure detection in neonatal EEGRobert Hogan0Sean R. Mathieson1Aurel Luca2Soraia Ventura3Sean Griffin4Geraldine B. Boylan5John M. O’Toole6CergenX LtdCergenX LtdCergenX LtdCergenX LtdCergenX LtdCergenX LtdCergenX LtdAbstract Neonatal seizures require urgent treatment, but often go undetected without expert EEG monitoring. We have developed and validated a seizure detection model using retrospective EEG data from 332 neonates. A convolutional neural network was trained and tested on over 50,000 hours (n = 202) of annotated single-channel EEG containing 12,402 seizure events. This model was then validated on two independent multi-reviewer datasets (n = 51 and n = 79). Increasing data and model size improved performance: Matthews correlation coefficient (MCC) and Pearson’s correlation (r) increased by up to 50% (15%) with data (model) scaling. The largest model (21m parameters) achieved state-of-the-art on an open-access dataset (MCC = 0.764, r = 0.824, and AUC = 0.982). This model also attained expert-level performance on both validation sets, a first in this field, with no significant difference in inter-rater agreement when the model replaces an expert (∣Δ κ∣ < 0.094, p > 0.05).https://doi.org/10.1038/s41746-024-01416-x |
| spellingShingle | Robert Hogan Sean R. Mathieson Aurel Luca Soraia Ventura Sean Griffin Geraldine B. Boylan John M. O’Toole Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG npj Digital Medicine |
| title | Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG |
| title_full | Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG |
| title_fullStr | Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG |
| title_full_unstemmed | Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG |
| title_short | Scaling convolutional neural networks achieves expert level seizure detection in neonatal EEG |
| title_sort | scaling convolutional neural networks achieves expert level seizure detection in neonatal eeg |
| url | https://doi.org/10.1038/s41746-024-01416-x |
| work_keys_str_mv | AT roberthogan scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg AT seanrmathieson scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg AT aurelluca scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg AT soraiaventura scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg AT seangriffin scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg AT geraldinebboylan scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg AT johnmotoole scalingconvolutionalneuralnetworksachievesexpertlevelseizuredetectioninneonataleeg |