Metabolism of the sleeping brain: potential links between sleep microarchitecture and peripheral blood glucose

Blood glucose dysregulation is a hallmark of diabetes, a major driver of excess morbidity and mortality [1]. Extensive observational, cross-sectional, longitudinal, and experimental data have established insufficient and disrupted sleep as significant risk factors for diabetes [2, 3]. Notably, experimental studies show that even short-term sleep restriction in healthy adults rapidly impairs insulin sensitivity [3, 4], resulting in blood glucose dysregulation—highlighting a potential mechanism linking poor sleep and diabetes risk [5]. Despite such data, direct links between sleep at the neurophysiological level and peripheral regulation of blood glucose are poorly understood. This knowledge gap contributes to the lack of effective, evidence-based interventions to mitigate sleep-related diabetes risk [5].

Polysomnography, the gold standard of quantifying sleep, may hold the potential to advance our understanding of the links between the sleeping brain and peripheral glucose control. For example, experimental slow-wave sleep (SWS) suppression in humans impairs insulin sensitivity and glucose homeostasis [6–8], suggesting a potential role for SWS in peripheral glucose regulation. Yet, little is known about the potential links between sleep microarchitecture—such as sleep spindles, slow oscillations, and sharp-wave ripples—and glucose metabolism, as these neurophysiological markers of plasticity during sleep have rather been associated with memory consolidation [9]. Animal studies provide intriguing evidence in this area; for instance, hippocampal sharp-wave ripples in rats precede decreases in peripheral glucose levels, with optogenetic manipulation suggesting a causal link [10]. Yet, whether similar associations exist in humans remains largely unexplored.

In this current issue of SLEEP, Yang and colleagues [11] build on these observations in rats by examining links between sleep microarchitecture and peripheral glucose in humans. This research examined sleep macro- and microarchitecture using sleep electroencephalography (EEG), and peripheral glucose levels during sleep using a unique application of continuous glucose monitors (CGMs) that quantify interstitial glucose concentrations. As surface EEG in humans cannot directly quantify hippocampal ripples, the authors focused on spindles and slow oscillations—two surface electrophysiological markers that play a key role in sleep-dependent memory processing [11, 12].

The final sample consisted of 10 participants (five women) with two consecutive nights of data each. The main findings revealed small but significant glucose changes time-locked to oscillatory events involved in plasticity during sleep. Specifically, glucose levels decreased (~0.24 mg/dL) within 1–6 min after sleep spindles, mirroring prior findings in rats, though with a smaller effect size [10]. Alternatively, relatively small but significant increases in glucose (~0.51 mg/dL) were detected about 5–11 min following slow oscillations. By focusing analyses on the coupling between spindles and slow oscillations, the authors found that the positive association between slow oscillations and glucose largely outweighed the negative association between spindle events and glucose. Further analyses showed that awakenings and microarousals were linked to rapid increases in glucose, and transitions into rapid eye movement (REM) sleep (lasting longer than 12) were followed by decreases in glucose about 10–14 min later. Importantly, although slow-oscillation density and slow-oscillation-spindle events were higher on the second versus first night in the lab, suggesting a slight first-night effect, the main findings were similar across nights.

These results provide the first evidence in humans linking sleep spindles and slow oscillations to peripheral glucose dynamics on a minute-to-minute timescale. As the authors suggest, the observed divergence—glucose decreasing after spindles but increasing after slow oscillations—may reflect different underlying neurophysiological processes. Spindles are strongly linked to memory consolidation, whereas slow oscillations may serve broader functions [13]. However, the study was not designed to determine causal mechanisms, raising new questions about whether these glucose changes result from altered neuronal activity, shifts in energy demand, autonomic modulation, or other potential mechanisms. Despite such possibilities, given the technical and novel aspect of the current findings, additional trials are needed to replicate the data and test specific hypothesized mechanisms underlying the reported associations.

Regarding future research aimed at replicating and building on the reported findings [11], there are some important considerations. First, CGMs measure interstitial glucose rather than blood glucose, with an approximate 5–6-min delay in changes between the blood and interstitial space. It is therefore possible that decreases in glucose at 1–6 min following spindles may actually occur prior to, rather than after, the spindles. This could suggest that the spindle itself is not driving the change in glucose, but both the spindle and the change in glucose may be the consequence of another interrelated factor (e.g. decreased noradrenergic locus coeruleus activity as speculated by the authors [11]). This raises a technical challenge for future research, although blood samples as frequent as every 60 s can be collected through intravenous catheters during wakefulness, this is more challenging during sleep without waking the participant. Second, the detected changes in glucose were physiologically small (< 1 mg/dl), raising questions about their clinical significance, particularly in relation to diabetes risk. Prior studies linking sleep restriction to diabetes have emphasized insulin sensitivity [5] rather than minor glucose fluctuations. Conducting similar research in populations with prediabetes or diabetes, or under controlled glucose manipulation, could provide greater insight into the metabolic implications. Such work is broadly important for the field because in some sleep restriction studies, glucose dysregulation still occurs when there is a primary loss of REM sleep and preservation of SWS [3], suggesting that SWS is not the sole sleep stage linked to glucose regulation. Studies of increasing nightly sleep duration to improve glucose dysregulation are also emerging [5], and such an experimental paradigm could help shed light on the newly reported associations between microarchitecture and glucose [11]. Third, although results from Yang and colleagues were consistent across frontal, central, and parietal electrodes, the magnitudes of the glucose changes were statistically larger for more posterior regions. The relevance of these regional differences warrants attention in future research given that features of sleep spindles and slow oscillations also exhibit topographical variation. Implementing high-density EEG is one strategy to help advance our understanding of the physiological significance of the observed topographical variation. Fourth, and as noted by the authors [11], future research can employ experimental interventions—such as acoustic stimulation time-locked to specific electrophysiological events of interest—to reveal a causal link between sleep features and glucose levels. Fifth, more sophisticated analyses of coupling events, incorporating the rhythm of the oscillations on the whole [14], may provide additional insights into the mechanisms underlying the associations with glucose. Lastly, as the reported oscillation events are linked to sleep-dependent memory consolidation, adding memory tasks to this line of inquiry could provide a more complete and integrated assessment of the potential links between EEG microarchitecture, memory consolidation, and glucose metabolism during sleep.

These exciting findings [11] from Yang and colleagues open a new door for the field to focus on the links between the sleeping brain and peripheral glucose control. Unraveling the mechanisms underlying the identified associations has potential wide-ranging applicability from sleep-dependent memory processing to cardiometabolic and cognitive health. Although replication is needed in larger studies, future research incorporating targeted closed-loop manipulations of sleep architecture, neuroimaging, metabolic assessments, and memory tasks will be essential to unravel the physiological significance of these complex associations. In aggregate, understanding the complete physiology of these observations in healthy adults could then help the field better understand what physiological dysregulation may be contributing to related disease states.

Disclosure statement

Financial Disclosure: C.M.D. reports funding from the National Institutes of Health and the Ben B. and Iris M. Margolis Foundation that is unrelated to this work. Non-Financial Disclosure: None.

Data availability

No new data were generated or analyzed in support of this work.