Background

Sleep and wakefulness are actively controlled by the interplay of distinct neural circuits in the brain (Weber and Dan, 2016; Scammell et al., 2017). Centered on the mutual inhibition between sleep- and wakefulness-promoting circuits, a flip-flop model has been proposed to explain the mechanisms of sleep-wake state transitions (Saper et al., 2001; Saper et al., 2010). The discovery of the ascending reticular activating system (Moruzzi and Magoun, 1949), the orexin neurons (de Lecea et al., 1998; Sakurai et al., 1998), and further studies of neuromodulatory systems have greatly advanced our understanding of the neural circuits supporting wakefulness (Brown et al., 2012; Lee and Dan, 2012; Scammell et al., 2017). By contrast, the neural circuits controlling sleep have remained elusive. Several sleep-promoting regions have been identified, including the preoptic area (POA, particularly the ventrolateral preoptic area and median preoptic nucleus, or VLPO and MPO, respectively) (Sherin et al., 1996; John and Kumar, 1998; Lu et al., 2000; Alam et al., 2014; Kroeger et al., 2018), the parafacial zone (Anaclet et al., 2012 and 2014), and the basal forebrain (Xu et al., 2015). However, whether these brain structures are involved in sleep initiation or maintenance remains largely unknown. GABAergic neurons in these brain regions are sleep-active, and their activation promotes non-rapid eye movement (NREM) sleep (Scammell et al., 2017). In addition to the role of GABAergic neurons in sleep regulation, recent studies have identified glutamatergic neurons that promote NREM sleep in a few brain areas, including the perioculomotor region of the midbrain (Zhang et al., 2019), the ventrolateral periaqueductal gray of the midbrain (Zhong et al., 2019), and the posterior thalamus (Ma et al., 2019).

The POA is the most intensively studied sleep-active and sleep-promoting region. Immunohistochemistry studies in rats showed the existence of sleep-active GABAergic neurons in the VLPO and a positive correlation between the number of c-Fos-positive cells and the amount of NREM sleep (Sherin et al., 1996; Lu et al., 2002). Consistently, electrophysiological recordings demonstrated that the firing rate of VLPO sleep-active neurons correlates with the depth of NREM sleep [indicated by electroencephalogram (EEG) delta power] (Szymusiak et al., 1998; Alam et al., 2014). Notably, VLPO sleep-active neurons displayed lower activity in the wake-sleep transition period than in the subsequent sleep episode (discharge rates further increased significantly from light to deep NREM sleep) (Szymusiak et al., 1998). These findings suggest that VLPO neurons might be involved in sleep maintenance, whereas other neurons may be responsible for sleep initiation. Using c-Fos activity approach, retrograde tracing, calcium imaging, and optogenetic and chemogenetic manipulations, we recently identified a population of POA-projecting glutamatergic neurons in the ventrolateral medulla (VLM) that controls the transitions from wakefulness to NREM sleep (Teng et al., 2022).

Here, we describe a detailed protocol for deep brain calcium imaging in the brainstem with simultaneous EEG/electromyography (EMG) recording to monitor neuronal activity during wake and sleep cycles (Figure 1). Compared to fiber photometry recording, this microendoscopic calcium imaging technique provides single-cell resolution to distinguish the function of subpopulations of VLM neurons in behaving mouse. Compared to in vivo electrophysiological recording, this imaging method allows the examination of neuronal activity in genetically defined cells. Similar in vivo calcium imaging techniques have been used to study sleep-related neuronal activity in other brain regions, including galanin-expressing GABAergic neurons in the dorsomedial hypothalamus (Chen et al., 2018) and neurotensinergic neurons in the midbrain (Zhong et al., 2019). A detailed imaging protocol in the cortical and subcortical areas has been published (Resendez et al., 2016). However, imaging the brainstem in freely moving animals typically involves significant movement artifacts. To overcome this issue, a previous study used anchoring tungsten wires glued to the gradient refractive index (GRIN) lens to stabilize the field of view (Gong et al., 2020). Here, we adopted this method to image VLM neurons in wake and sleep cycles. The whole experimental procedure takes approximately two months (Figure 1A). Our imaging protocol should be applicable to other deep brain regions that are involved in sleep regulation.

Figure 1. Imaging medulla glutamatergic neurons during sleep. A. Timeline of experimental design. B. Top: schematic of microendoscopic calcium imaging in the brainstem. Tungsten wires attached to the gradient refractive index (GRIN) lens were used to stabilize the field of view. Bottom: averaged GCaMP fluorescence image in the field of view and activity map of a representative imaging session in a Vglut2-Cre mouse. Scale bar, 50 μm. C. Example data showing electroencephalogram (EEG), electromyography (EMG), and representative calcium traces (DF/F) in an imaging session. Bottom: enlarged window showing calcium activity in three cells.