Dynamic electrocortical states and paradoxical complexity during desflurane anesthesia

Background: How general anesthesia alters the dynamics of electrocortical activity is crucial to understand the neural mechanisms of unconsciousness. Local cortical activity undergoes spontaneous transitions at constant anesthetic concentration. The spatial organization and temporal dynamics of state transitions in large-scale electrocortical activity is incompletely understood. Methods: Epidural electrocorticogram was recorded from the right hemisphere in 8 rats (14 experiments) using chronically implanted 32-channel flexible electrode arrays during desflurane anesthesia at 6%, 4%, 2%, 0% inhaled concentrations, each maintained for 1 hour. Cortical states were identified by principal component analysis of power spectrograms followed by density-based clustering simultaneously across all anesthetic conditions. State-specific spatiotemporal complexity was quantified by the normalized Lempel-Ziv algorithm to capture signal variability beyond spectral effects. Temporal dynamics were assessed by state occurrence, dwell times, and transition probabilities. Results: Seven cortical states were identified. Six states generally tracked anesthetic depth with an increase in delta power and decrease in complexity, but their occurrence was not tied to any anesthetic level (p<0.001). The 7th state was a paradoxical, activated state that mostly occurred during deep anesthesia and was marked by reduced delta (p<0.001) and elevated complexity (p<0.001). Cortical activity was more likely to remain in a given state than to switch (mean dwell time of 136.55 s, persistence likelihood of 99.36%). When transitions occurred, they followed structured, non-random dynamics, primarily within light- or deep-anesthesia states (FDR-corrected p<0.05), with a mild tendency to exit deep states and enter light states (p=0.0039), consistent with anesthetic emergence. Conclusions: Electrocortical activity is not a unitary function of anesthetic concentration but involves spontaneous dynamics and paradoxically activated states with increased global complexity during deep anesthesia. The results suggest ongoing spontaneous reorganization of cortical activity during prolonged anesthetic challenge and provide new insight into anesthesia-induced brain dynamics that may inform future strategies for monitoring and manipulating the state of consciousness and facilitating recovery from general anesthesia.