Are There Common Network-level Correlates of the Anesthetized Brain in Infants and Adults?

Anesthetic care of infants and children has evolved considerably over the last four decades. In the mid-1980s, infants were undergoing invasive procedures with either no or “light” anesthesia supplemented with a neuromuscular blocking agent. Parents became appalled, which motivated a public outcry and sparked an intense discussion among anesthesia providers on how best to anesthetize infants.¹  This led to a paradigm shift in pediatric anesthesiology, with providers recognizing the possibility and consequences of pain in infants undergoing surgery and reconsidering anesthetic regimens accordingly. The pendulum has now swung in the opposite direction, as the conversation in pediatric anesthesia over the last decade has been dominated by neurotoxicity and the potentially detrimental effects of anesthetics in the developing brain, brought into public view by a recent Food and Drug Administration black box warning on anesthetic medications for young children and pregnant women. As every pediatric anesthesia provider will attest, infants and children are not just “little adults,” yet in accomplishing one of the fundamental goals of anesthetic care, methods of titrating anesthetics in infants are based on the same variables used in adult patients (i.e., minimum alveolar concentration values, patient movement, and hemodynamic changes). This is, in part, due to an incomplete understanding of the neurobiologic mechanisms of anesthesia and lack of surrogate markers of unconsciousness in the developing brain. In this issue of Anesthesiology, Pappas et al.²  present an elegant study that begins to inform this current gap in knowledge.

The study was a retrospective cohort analysis of multichannel electroencephalographic recordings of 20 infants, aged 0 to 3.9 months old, who underwent sevoflurane anesthesia for elective surgical procedures. The goals of the study were to identify changes in functional connectivity and network properties in the infant brain during two time points: the maintenance phase (end-tidal sevoflurane, 1.8 to 2.5%) and emergence (end-tidal sevoflurane, 0 to 0.3%). First, the investigators used a measure of functional connectivity, i.e., the statistical covariation of activities across different brain regions. Their electroencephalographic analysis was limited to the slow-wave or δ bandwidth (1 to 4 Hz) and analyzed in two ways: (1) directly from the electroencephalographic electrode/recording location (sensor) and (2) a mathematical transformation of the electroencephalographic data that estimates the neural origin of the signal in the brain (source). Both analyses yielded similar results. During the maintenance phase, functional connectivity was found to be decreased across several locations, including the frontal–parietal region. The breakdown of frontal–parietal connectivity has been postulated to be a neural correlate of unconsciousness and mechanism of disrupted long-range information transfer in adults,³  although this connectivity pattern is dynamic during anesthesia.^4,5  The authors interpreted the finding of functional cortical disconnection as a potential mechanism for the loss of consciousness in infants. Additionally, Pappas et al.²  used network science methods to explore how infant brain network architecture (in this case, modularity, or the degree to which functional networks are characterized by individual modules or an integrated whole) and whole-brain dynamic patterns (complexity) changed during these two periods in the δ frequency band. They found that brain networks were more fragmented and less complex during general anesthesia, supporting the hypothesis of impaired integration and a constrained repertoire of states.

These findings are striking because this emerging network-level picture of general anesthesia in infants is remarkably consistent with that of adults, despite substantial differences in the properties of the electroencephalogram. Collectively, the past studies of anesthetized adults and these new data in anesthetized infants suggest the remarkable possibility that there might be a common network-level mechanism and marker of anesthetic-induced unconsciousness that is age-invariant. This heralds the possibility of a monitor that, in contrast to current commercial options, is effective even in the very young. The finding that networks appear to be less integrated during general anesthesia is also consistent with leading theories of consciousness.⁶ 

In adults, hypercoherent frontal α oscillations,⁷  breakdown in functional brain connectivity,⁸  reduced complexity in brain oscillatory patterns, and changes in network properties⁸  have all been consistently identified. In infants undergoing general anesthesia, there is a lack of clear electroencephalographic markers of unconsciousness, and the relevance of adult connectivity and network-based markers has been previously unexplored. To highlight the differences between infants and adults, figure 1 shows example spectrograms (i.e., plots of electroencephalogram frequency by time, with color indicating power) and coherograms (i.e., plots of synchrony between two signals at the same frequency, with color indicating strength) during the maintenance phase of sevoflurane anesthesia. Note, for example, the lack of coherent frontal α oscillations in the infant,^9,10  which can be clearly identified in adults during anesthesia. Additionally, in figure 2A presented in the article by Pappas et al.,²  notice the striking similarity between maintenance and emergence from general anesthesia in infants. It is remarkable that, despite these differences in infants and adults, common patterns emerge when examining the network level.

Although these results are exciting, the study was relatively small, with the possibility of anesthetic and surgical stimulus heterogeneity. Additionally, although slow waves are the most prominent brain oscillation during this developmental period, additional frequency bands should be characterized in the future, as should different recording time points (such as baseline measurements) and different ages or stages of development. Finally, the possibility of dynamic transitions in these connectivity patterns, as has been shown in adults,^4,5  should be explored.

In conclusion, the study by Pappas et al.²  shows us that general anesthetics alter functional brain networks in infants in a way that is similar to adults. This establishes an important foundation for future work that can systematically identify age-invariant markers of anesthetic-induced consciousness.