Related Experiment Video
Updated: Jun 17, 2026
Application of an Amplitude-integrated EEG Monitor (Cerebral Function Monitor) to Neonates
Published on: September 6, 2017
Automatic classification of background EEG activity in healthy and sick neonates
Johan Löfhede1, Magnus Thordstein, Nils Löfgren
1School of Engineering, University College of Borås, Borås, Sweden. johan.lofhede@hb.se
Insights
This study introduces an automated system for classifying neonatal electroencephalogram (EEG) activity, accurately distinguishing pathological burst suppression from healthy sleep states in newborns. The system also segments burst suppression patterns for critical parameter calculation.
Area of Science:
- Neonatal neurology
- Biomedical signal processing
- Machine learning in healthcare
Background:
- Neonatal intensive care units (NICUs) require sophisticated monitoring systems.
- Differentiating normal neonatal electroencephalogram (EEG) from pathological patterns is crucial for patient care.
- Existing methods for EEG analysis in newborns can be labor-intensive and subjective.
Purpose of the Study:
- To develop an automated classification scheme for background EEG activity in newborn infants.
- To differentiate pathological burst suppression patterns from normal sleep states in neonates.
- To enable the calculation of clinically relevant parameters from burst suppression patterns.
Main Methods:
- Collected EEG data from 20 healthy full-term newborns across four behavioral states (active awake, quiet awake, active sleep, quiet sleep).
- Collected EEG data from 6 full-term newborns exhibiting burst suppression due to birth asphyxia.
- Utilized feature extraction from EEG signals combined with Fisher's linear discriminant classifier for automated classification.
- Developed a segmentation method to analyze burst and suppression phases within pathological EEG patterns.
Main Results:
- Achieved 100% classification accuracy in distinguishing burst suppression EEG from all four healthy neonatal EEG states.
- Reached 93% true positive classification for identifying quiet sleep EEG among healthy states.
- Successfully segmented burst suppression EEG into burst and suppression phases with approximately 4% error rate.
- Enabled calculation of parameters like suppression length and burst suppression ratio.
Conclusions:
- The developed automated EEG classification system effectively identifies pathological burst suppression in newborns.
- The system provides accurate segmentation of burst suppression patterns, facilitating the calculation of key clinical parameters.
- This technology holds promise for improving monitoring and diagnosis in neonatal intensive care units.
Abstract:
The overall aim of our research is to develop methods for a monitoring system to be used at neonatal intensive care units. When monitoring a baby, a range of different types of background activity needs to be considered. In this work, we have developed a scheme for automatic classification of background EEG activity in newborn babies. EEG from six full-term babies who were displaying a burst suppression pattern while suffering from the after-effects of asphyxia during birth was included along with EEG from 20 full-term healthy newborn babies. The signals from the healthy babies were divided into four behavioural states: active awake, quiet awake, active sleep and quiet sleep. By using a number of features extracted from the EEG together with Fisher's linear discriminant classifier we have managed to achieve 100% correct classification when separating burst suppression EEG from all four healthy EEG types and 93% true positive classification when separating quiet sleep from the other types. The other three sleep stages could not be classified. When the pathological burst suppression pattern was detected, the analysis was taken one step further and the signal was segmented into burst and suppression, allowing clinically relevant parameters such as suppression length and burst suppression ratio to be calculated. The segmentation of the burst suppression EEG works well, with a probability of error around 4%.
