Cardiac Signals Show Stronger Directional Influence on Infant Brain Activity in CHD
Sleep recordings from infants and toddlers with congenital heart disease point to a measurable shift in how cardiac autonomic activity and cortical rhythms interact. Using simultaneous electrocardiography and electroencephalography during sleep, investigators examined whether directionality between heart-rate variability and EEG band power differed between children with CHD and healthy controls. The cohort was small but carefully paired for signal analysis: 14 infants and toddlers with heterogeneous CHD diagnoses and 16 healthy controls.
The central finding was not a complete reversal of brain-heart communication. Both groups showed predominantly positive net directionality indexes, indicating that cardiac autonomic signals were generally stronger in the modeled direction than the reverse. The difference emerged in the size of that directional signal. Children with CHD demonstrated higher net cardiac-to-cortical estimates for LF/HF–alpha and LF/HF–beta after permutation-based false discovery rate control.
Directionality Measures Suggest a Shift in Autonomic-Cortical Coupling
The analysis relied on temporal alignment of heart-rate variability features with EEG band-power features, then applied bidirectional symbolic transfer entropy based on amplitude-bin quantization. In practical terms, that method asks whether fluctuations in one biological signal improve prediction of future fluctuations in another. A higher net directionality index toward the cortex implies that cardiac autonomic dynamics carried more predictive weight for cortical activity than the reverse, at least within the structure of these sleep recordings.
One signal stood out from the rest.
The higher LF/HF–alpha and LF/HF–beta estimates in the CHD group are biologically plausible. The LF/HF ratio is often used as a rough marker of autonomic balance, and alpha and beta EEG bands are sensitive to changing arousal and cortical state. During sleep, those systems are not static; they remain coupled through respiratory, autonomic, and neurophysiological loops that can be altered when congenital cardiac anatomy changes oxygen delivery, circulation, or autonomic regulation early in life. A cardiac lesion does not have to injure the brain directly to reshape this coupling. Chronic differences in hemodynamic load, oxygenation, and autonomic compensation may leave a detectable imprint on cortical oscillatory behavior.
Heterogeneity Limits Any Single CHD Signature
The cohort’s breadth is also its weakness. The CHD group included heterogeneous diagnoses, which means the observed signal cannot be collapsed into a single disease phenotype. Different cardiac malformations can produce distinct patterns of hemodynamic stress, perioperative exposure, and developmental vulnerability. A broad group-level effect may therefore reflect overlapping biology rather than one uniform mechanism.
Interpretation is further constrained by acquisition-setting differences and unresolved effects of the underlying feature statistics. Those caveats matter because transfer-entropy methods are sensitive to preprocessing choices, feature binning, and the statistical behavior of the time series itself. In a small sample, subtle analytic decisions can influence whether a directional effect reaches significance. The result is best viewed as exploratory physiology rather than a bedside biomarker.
Why the Sleep State Matters for Infant Neurocardiac Physiology
Sleep offers a useful window into early brain-heart physiology because it reduces some of the noise of wakeful behavior while preserving meaningful autonomic and cortical dynamics. In infants and toddlers, sleep EEG is still shaped by developmental maturation, and cardiac variability remains highly dynamic. If CHD shifts the balance of autonomic drive, sleep may be one of the first settings where that imbalance becomes visible through signal analysis.
The choice to examine LF/HF ratio alongside alpha and beta power is not accidental. These domains sit at the intersection of autonomic regulation and cortical activation. Alpha activity often tracks state-dependent cortical inhibition, while beta activity can reflect more alert or active neural processing. When cardiac autonomic influence rises relative to cortical activity in these bands, the finding may point to altered developmental coupling rather than a single disease-specific biomarker.
Prognosis Cannot Be Read Into This Dataset
This study does not establish individual clinical risk, altered physiology in every child with CHD, or a basis for prognostic use. The authors explicitly avoid that leap, and the data support restraint. A group-level difference in directional coupling is not the same as a validated predictor of neurodevelopmental outcome, perioperative risk, or later cognitive performance.
Longitudinal follow-up would be the next meaningful step. Repeating these recordings across development could show whether the directional signal persists, widens, normalizes, or varies by lesion type and treatment course. Methodological validation will matter just as much as biological replication, because symbolic transfer entropy must prove stable before it can be used beyond exploratory analysis.
For now, the paper adds a useful physiologic observation: early-life CHD may be associated with stronger cardiac-to-cortical statistical influence during sleep, especially across LF/HF-linked alpha and beta features. The study is too small and too heterogeneous to define a clinical phenotype, but it gives researchers a more precise target for future work on the infant neurocardiac axis.