Skip to main content
Log in

Pilot study of a single-channel EEG seizure detection algorithm using machine learning

  • Original Article
  • Published:
Child's Nervous System Aims and scope Submit manuscript

Abstract

Objective

Seizures are one of the most common emergencies in the neonatal intensive care unit (NICU). They are identified through visual inspection of electroencephalography (EEG) reports and treated by neurophysiologic experts. To support clinical seizure detection, several feature-based automatic neonatal seizure detection algorithms have been proposed. However, as they were unsuitable for clinical application due to their low accuracy, we developed a new seizure detection algorithm using machine learning for single-channel EEG to overcome these limitations.

Methods

The dataset applied in our algorithm contains EEG recordings from human neonates. A 19-channel EEG system recorded the brain waves of 79 term neonates admitted to the NICU at the Helsinki University Hospital. From these datasets, we selected six patients with conformational seizure annotations for the pilot study and allocated four and two patients for our training and testing datasets, respectively. The presence of seizures in the EEGs was annotated independently by three experts through visual interpretation. We divided the data into epochs of 5 s each and further defined a seizure block to label the annotations from each expert recorded every second. Subsequently, to create a balanced dataset, any data point with a non-seizure label was moved to the training and test dataset.

Result

The developed principal component feature–extracted machine learning algorithm used 62.5% of the relative time (only 5 s for decision) of the baseline, reaching an area under the ROC curve score of 0.91. The effect of diversified parameters was meticulously examined, and 100 principal components were extracted to optimize the model performance.

Conclusion

Our machine learning–based seizure detection algorithm exhibited the potential for clinical application in NICUs, general wards, and at home and proved its convenience by requiring only a single channel for implementation.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

References

  1. Rosen JB, Hamerman E, Sitcoske M, Glowa JR, Schulkin JJBN (1996) Hyperexcitability: exaggerated fear-potentiated startle produced by partial amygdala kindling. 110:43

    CAS  Google Scholar 

  2. Buzsaki G, Ponomareff G, Bayardo F, Ruiz R (1989) Gage FJN. Neuronal activity in the subcortically denervated hippocampus: a chronic model for epilepsy. 28:527–538

    CAS  Google Scholar 

  3. (!!! INVALID CITATION !!! 7,12,18))

  4. Boylan G, Burgoyne L, Moore C, O’Flaherty B (2010) Rennie JJAp. An international survey of EEG use in the neonatal intensive care unit. 99:1150–1155

    Google Scholar 

  5. Glass HC, Glidden D, Jeremy RJ, Barkovich AJ, Ferriero DM (2009) Miller SPJTJop. Clinical neonatal seizures are independently associated with outcome in infants at risk for hypoxic-ischemic brain injury. 155:318–323

    Google Scholar 

  6. Wirrell EC, Armstrong EA, Osman LD, Yager JYJPr (2001) Prolonged seizures exacerbate perinatal hypoxic-ischemic brain damage. 50: 445

  7. Malone A, Anthony Ryan C, Fitzgerald A, Burgoyne L, Connolly S (2009) Boylan GBJE. Interobserver agreement in neonatal seizure identification. 50:2097–2101

    Google Scholar 

  8. Azzopardi D, Strohm B, Edwards AD, Halliday H, Juszczak E, Levene M, Thoresen M, Whitelaw A (2009) Brocklehurst PJAoDiC-F, Edition N. Treatment of asphyxiated newborns with moderate hypothermia in routine clinical practice: how cooling is managed in the UK outside a clinical trial. 94:F260–F264

    CAS  Google Scholar 

  9. Toet MC (2009) Lemmers PMJEhd. Brain monitoring in neonates. 85:77–84

    Google Scholar 

  10. Rennie J, Chorley G, Boylan G, Pressler R, Nguyen Y (2004) Hooper RJAoDiC-F, Edition N. Non-expert use of the cerebral function monitor for neonatal seizure detection. 89:F37–F40

    CAS  Google Scholar 

  11. Shellhaas RA, Soaita AI (2007) Clancy RRJP. Sensitivity of amplitude-integrated electroencephalography for neonatal seizure detection. 120:770–777

    Google Scholar 

  12. Evans E, Koh S, Lerner J, Sankar R (2010) Garg MJAoDiC-F, Edition N. Accuracy of amplitude integrated EEG in a neonatal cohort. 95:F169–F173

    CAS  Google Scholar 

  13. Boubchir L, Daachi B, Pangracious V (2017) A review of feature extraction for EEG epileptic seizure detection and classification. 2017 40th International Conference on Telecommunications and Signal Processing (TSP). IEEE, pp 456-460

  14. Ansari AH, Cherian PJ, Caicedo A, Naulaers G, De Vos M, Van Huffel SJIjons (2019) Neonatal seizure detection using deep convolutional neural networks. 29: 1850011

  15. Lu Y, Ma Y, Chen C, Wang YJT (2018) Care H. Classification of single-channel EEG signals for epileptic seizures detection based on hybrid features. 26:337–346

    Google Scholar 

  16. Stevenson N, Tapani K, Lauronen L, Vanhatalo SJSd (2019) A dataset of neonatal EEG recordings with seizure annotations. 6: 190039

  17. Song F, Guo Z, Mei D (2010) Feature selection using principal component analysis. 2010 international conference on system science, engineering design and manufacturing informatization. IEEE, pp 27-30

  18. Faul S, Temko A, Marnane W (2009) Age-independent seizure detection. 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE, pp 6612-6615

  19. O’Shea A, Lightbody G, Boylan G (2020) Temko AJNN. Neonatal seizure detection from raw multi-channel EEG using a fully convolutional architecture. 123:12–25

    Google Scholar 

  20. Rahman MM, Davis DNJIJoML, Computing (2013) Addressing the class imbalance problem in medical datasets. 3: 224

  21. Lachaux JP, Rodriguez E, Martinerie J (1999) Varela FJJHbm. Measuring phase synchrony in brain signals. 8:194–208

    CAS  Google Scholar 

  22. Van Dyk DA (2001) Meng X-LJJoC, Statistics G. The art of data augmentation. 10:1–50

    Google Scholar 

  23. Cubuk ED, Zoph B, Shlens J, Le QV (2020) Randaugment: practical automated data augmentation with a reduced search space. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, pp 702-703

  24. Wang T, Qin Z, Jin Z, Zhang SJJoS, Software (2010) Handling over-fitting in test cost-sensitive decision tree learning by feature selection, smoothing and pruning. 83: 1137-1147

  25. Cawley GC (2007) Talbot NLJJoMLR. Preventing over-fitting during model selection via Bayesian regularisation of the hyper-parameters. 8:841–861

    Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Dong-Seok Kim.

Ethics declarations

Conflict of interest

On behalf of all authors, the corresponding author states that there is no conflict of interest.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Ryu, S., Back, S., Lee, S. et al. Pilot study of a single-channel EEG seizure detection algorithm using machine learning. Childs Nerv Syst 37, 2239–2244 (2021). https://doi.org/10.1007/s00381-020-05011-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00381-020-05011-9

Keywords

Navigation