Arousal neurons that anticipate deviations in blood glucose.

The brain needs to track body energy state to optimize physiology and behavior. Important information about current and future energy states is contained in minute-to-minute fluctuations in blood glucose. However, it is unclear whether brain glucose sensors are capable of responding to this temporal structure to extract such information. In behaving mammals, hypothalamic hypocretin/orexin neurons (HONs) control arousal and are proposed to sense energy balance, yet recent studies show that HON activity varies rapidly (over seconds) with locomotion. It remains unknown if and how HONs reflect blood glucose fluctuations. Here, for the first time, we co-monitored HON activity and blood glucose in behaving mice, uncovering inhibition of HONs by rising blood glucose. Surprisingly, peak HON signals (population waves) anticipated peak glucose deviations by several minutes. In multivariate analysis of diverse HON activity correlates (locomotion, metabolic variables), the glucose temporal gradient (negative first derivative) emerged as prime predictor of HON population activity on the minute timescales, thereby explaining this anticipatory response. Furthermore, 2-photon imaging of >900 individual HONs revealed parallel communication of absolute blood glucose values and their derivatives in distinct HON subsets. Thus, HONs transmit multiplexed slowly and rapidly changing information, and, in the slow bandwidth, extract temporal features from blood glucose dynamics that suggest a hybrid of anticipatory and proportional logic in brain responses to blood glucose. ### Competing Interest Statement The authors have declared no competing interest.