Two types of slow waves in anesthetized and sleeping brains

Deep sleep and anesthesia have contrasting effects on memory, yet at the microscopic scale they appear to produce similar neural network dynamics consisting of slow waves associated with alternating transients of high activity (UP states) and silence (DOWN states). Here, UP and DOWN state dynamics are analyzed in cortical recordings from human, monkey, and rat and found to be robustly different between deep sleep and anesthesia. We found that the temporal statistics of UP and DOWN states is robustly different in natural slow-wave sleep compared to the slow-waves of anesthesia. Computational models predict that an interplay between noise and spike-frequency adaptation can yield a transition from sleep-like to anesthesia-like slow-wave dynamics. These predictions were confirmed by pharmacological manipulations in vitro, inducing a switch between the two types of slow-waves. The results show that the strong adaptation found in anesthesia, but not in sleep, can produce a different type of slow oscillations, less sensitive to external inputs, that may prevent memory formation.