Cognitive Outcome Prediction in Infants With Neonatal Hypoxic-Ischemic Encephalopathy Based on Functional...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Developmental Pediatrics

Background:

• Neonatal hypoxic-ischemic encephalopathy (HIE) poses a significant risk for impaired neurodevelopmental outcomes, particularly cognitive deficits.
• Early identification of infants at risk for cognitive impairment is crucial for timely intervention and improved long-term prognosis.
• Current methods for predicting neurodevelopmental outcomes in HIE survivors have limitations.

Purpose of the Study:

• To investigate the utility of advanced quantitative electroencephalography (qEEG) analysis for early prediction of cognitive outcomes in infants with HIE.
• To explore the association between specific qEEG-derived features and cognitive development at two years of age.
• To develop and validate predictive models for cognitive impairment using neonatal qEEG data.

Main Methods:

• EEG data were collected within the first week of life from twenty infants diagnosed with HIE.
• Advanced qEEG analysis involved calculating graph-theoretical features (from weighted phase-lag index) and entropy metrics (sample entropy, permutation entropy, spectral entropy) within the noise-assisted multivariate empirical mode decomposition (NA-MEMD) domain.
• Regression models, including tree ensembles (boosted and bagged), were trained and tested using these qEEG features to predict cognitive scores at two years.

Main Results:

• Significant correlations were found in the delta band between neonatal qEEG graph attributes (radius, transitivity, global efficiency, characteristic path length) and entropy features (permutation entropy, spectral entropy) and cognitive development at two years.
• A boosted tree regression model utilizing entropy features achieved the highest prediction performance, with a root mean square error (RMSE) of 14.27, mean absolute error (MAE) of 12.07, and an R-squared value of 0.45.
• The findings indicate that specific qEEG features derived from neonatal recordings can reflect the early state of brain function.

Conclusions:

• The proposed qEEG features demonstrate potential as early predictive biomarkers for cognitive impairment in infants with HIE.
• These qEEG biomarkers could facilitate the identification of infants who would benefit from early, targeted neurodevelopmental interventions.
• This study highlights the value of advanced qEEG analysis in understanding and predicting neurodevelopmental trajectories after neonatal brain injury.