Schmidt Mellado, G. | Search bioRxiv

bioRxiv2023

Sensory event-related potential morphology predicts age in premature infants

Preterm infants undergo substantial neurosensory development in the first weeks after birth. Infants born prematurely are more likely to have long-term adverse neurological outcomes and early detection of abnormal brain development is essential for timely interventions. We investigated whether sensory-evoked cortical potentials could be used to accurately estimate the age of an infant. Such a model could be used to identify infants who deviate from normal neurodevelopment by comparing the brain age to the infants postmenstrual age (PMA). Infants aged between 28- and 40-weeks PMA from a training and test sample (consisting of 101 and 65 recording sessions in 82 and 14 infants, respectively) received trains of approximately 10 visual and 10 tactile stimuli (interstimulus interval approximately 10 seconds). PMA could be predicted accurately from the magnitude of the evoked responses (training set mean absolute error (MAE and 95% confidence intervals): 1.41 [1.14; 1.74] weeks, p = 0.0001; test set MAE: 1.55 [1.21; 1.95] weeks, p = 0.0002. Moreover, we show with two examples that brain age, and the deviations between brain age and PMA, may be biologically and clinically meaningful. By firstly demonstrating that brain age is correlated with a measure known to relate to maturity of the nervous system (based on animal and human literature, the magnitude of reflex withdrawal is used) and secondly by linking brain age to long-term neurological outcomes, we show that brain age deviations are related to biologically meaningful individual differences in the rate of functional nervous system maturation rather than noise generated by the model. In summary, we demonstrate that sensory-evoked potentials are predictive of age in premature infants. It takes less than 5 minutes to collect the stimulus electroencephalographic data required for our model, hence, increasing its potential utility in the busy neonatal care unit. This model could be used to detect abnormal development of infants response to sensory stimuli in their environment and may be predictive of later life abnormal neurodevelopmental outcome.

neuroscience↗

bioRxiv2023

Brain age as an estimator of neurodevelopmental outcome: A deep learning approach for neonatal cot-side monitoring

The preterm neonate can experience stressors that affect the rate of brain maturation and lead to long-term neurodevelopmental deficits. However, some neonates who are born early follow normal developmental trajectories. Extraction of data from electroencephalography (EEG) signals can be used to calculate the neonates brain age which can be compared to their true age. Discrepancies between true age and brain age (the brain age delta) can then be used to quantify maturational deviation, which has been shown to correlate with long-term abnormal neurodevelopmental outcomes. Nevertheless, current brain age models that are based on traditional analytical techniques are less suited to clinical cot-side monitoring due to their dependency on long-duration EEG recordings, the need to record activity across multiple EEG channels, and the manual calculation of predefined EEG features which is time-consuming and may not fully capture the wealth of information in the EEG signal. In this study, we propose an alternative deep-learning approach to determine brain age, which operates directly on the EEG, using a Convolutional Neural Network (CNN) block based on the Inception architecture (called Sinc). Using this deep-learning approach on a dataset of preterm infants with normal neurodevelopmental outcomes (where we assume brain age = postmenstrual age), we can calculate infant brain age with a Mean Absolute Error (MAE) of 0.78 weeks (equivalent to a brain age estimation error for the infant within +/- 5.5 days of their true age). Importantly, this level of accuracy can be achieved by recording only 20 minutes of EEG activity from a single channel. This compares favourably to the degree of accuracy that can be achieved using traditional methods that require long duration recordings (typically >2 hours of EEG activity) recorded from a higher density 8-electrode montage (MAE = 0.73 weeks). Importantly, the deep learning models brain age deltas also distinguish between neonates with normal and severely abnormal outcomes (Normal MAE = 0.71 weeks, severely abnormal MAE = 1.27 weeks, p=0.02, one-way ANOVA), making it highly suited for potential clinical applications. Lastly, in an independent dataset collected at an independent site, we demonstrate the models generalisability in age prediction, as accurate age predictions were also observed (MAE of 0.97 weeks). HighlightsO_LIPreterm stress exposure leads to long-term neurodevelopmental deficits C_LIO_LIDeficits are quantifiable using EEG-based brain age prediction errors C_LIO_LIOur deep-learning solution for brain age prediction outperforms previous approaches C_LIO_LIPredictions are achieved with only 20 mins EEG and a single bipolar channel C_LIO_LIPrediction errors correlate with long-term Bayley scale neurodevelopmental outcomes C_LI

neuroscience↗