Sensory Transmission is Bi-directionally Modulated by Astrocytic Ca 2+ in Barrel Cortex of Behaving Mice

: Different effects of astrocyte during sleep and awake have been extensively studied, especially for metabolic clearance by the glymphatic system, which works during sleep and stops working during waking states. However, how astrocytes contribute to modulation of sensory transmission during sleep and awake animals remain largely unknown. Recent advances in genetically encoded Ca 2+ indicators have provided a wealth of information on astrocytic Ca 2+ , especially in their fine perisynaptic processes, where astrocytic Ca 2+ most likely affects the synaptic function. Here we use two-photon microscopy to image astrocytic Ca 2+ signaling in freely moving mice trained to run on a wheel in combination with in vivo whole-cell recordings to evaluate the role of astrocytic Ca 2+ signaling in different behavior states. We found that there are two kinds of astrocytic Ca 2+ signaling: a small long-lasting Ca 2+ increase during sleep state and a sharp widespread but short-long-lasting Ca 2+ spike when the animal was awake (fluorescence increases were 23.2 ± 14.4% for whisker stimulation at sleep state, compared with 73.3 ± 11.7% for at awake state, paired t-test, p < 0.01) . The small Ca 2+ transients decreased extracellular K + , hyperpolarized the neurons, and suppressed sensory transmission; while the large Ca 2+ wave enhanced sensory input, contributing to reliable sensory transmission in aroused states. Locus coeruleus activation works as a switch between these two kinds of astrocytic Ca 2+ elevation. Thus, we show that cortical astrocytes play an important role in processing of sensory input. These two types of events appear to have different pharmacological sources and may play a different role in facilitating the efficacy of sensory transmission. release decreases during sleep, resulting in an expansion of the interstitial space and a subsequent potentiation of glymphatic-fluid transport. A recent study suggested that the astrocytic Ca 2+ elated with the glymphatic systems 3 , which suggested that neural slow waves are followed by hemodynamic oscillations, which in turn are coupled to CSF flow. These results demonstrate that the sleeping brain exhibits waves of CSF flow on a macroscopic scale, and these CSF dynamics are interlinked with neural and hemodynamic rhythms 3

Dynamic changes of astrocytic Ca 2+ during synaptic transmission have been extensively studied in the last decades [1][2][3][4][5][6] ; however, the topic remains controversial 1,7 . As stated in a recent review: "Few topics in neuroscience are as controversial as the idea that calcium concentration elevations in astrocytes release transmitters that regulate neuronal and vascular function" 1 . One reason might be that we were expecting logic input-output functions, but astrocytes are more complicated and a given input may have different output depending on brain states. For example, one study has demonstrated that the Ca 2+ elevation evoked by sensory stimulation can be separated into fast and slow components; fast Ca 2+ signals may initiate the cerebral blood flow response, while the slower and longer-lasting astrocytic Ca 2+ elevations may contribute to the sustained hemodynamic response 8 . We were expecting only one type of astrocytic Ca 2+ signaling: the induction is due to mGluR receptors 1 , their mechanisms are due to IP3R2 induced Ca 2+ release from endoplasm reticulum (ER) 9 , and the function is to release gliotransmitters [10][11][12][13][14] .
Different kinds of astrocytic Ca 2+ with different spatial location and temporal dynamics have started to be reported recently [15][16][17] . And new mechanisms involved in astrocytic Ca 2+ are reported, such as glutamate transporters 18 , Na + /Ca 2+ exchangers 19 , TRPA1 channels 1 . And new functions are appearing too 4 , such as increasing the fatty acid metabolism 18 , enhancing clearance of extracellular materials like the glymphatics 20,21 , affecting the circadian rhythm 4,22 .
Recently, the role of astrocyte in controlling global cortical network is a hot topic in many recent studies 4,23 , such as affecting the circadian rhythm 4,24 , controlling the slow oscillation 12 , affecting the neural circuits at different behavior states [25][26][27] . Previous experiments in our lab and other labs have found that the glymphatic system, the network of perivascular spaces through which cerebrospinal fluid and interstitial fluid can move through the brain [21][22] , is regulated by sleep and norepinephrine. In addition, it is reported that sensory input can induce local astrocytic Ca 2+ during anesthesia state [26][27] , whereas they can trigger widespread Ca 2+ signals in cerebral cortex mediated by alpha1 adrenergic receptor activation at waking states 28 . However, the local astrocytic Ca 2+ at the fine processes, especially at tripartites area, which got most of the Ca 2+ activity in an astrocyte, and the one most likely related to synaptic function 2 , are largely unknown. Recent advances in genetically encoded Ca 2+ indicators have provided a wealth of information on astrocytic Ca 2+ 7 . We here used genetically encoded Ca 2+ indicators (GCaMP6f) to monitor astrocytic Ca 2+ signaling in live free moving mice on a wheel in combination with dual whole-cell recordings in single neurons and local field potential (LFP) recordings in layer 2/3 of barrel cortex. We found that astrocytic Ca 2+ signaling can be differently induced and exert two different effects on sensory transmission at different behavior states. We also found that astrocytic Ca 2+ signaling have two opposing effects on sensory input: The small Ca 2+ increase in sleep mice suppresses sensory input, while the larger Ca 2+ transients enhances sensory input in alert awake mice. This is similar to the frequently described function of norepinephrine (NE) in increasing sensory input 29 . Thus, our study offers an explanation for the complex role of atsrocytes in sensory processing.

Astrocytic processes demonstrate two kinds of Ca 2+ signaling from sensory transmission
In order to probe into the astrocytic Ca 2+ signal at fine processes in sensory inputs in free moving mice, we used the genetically encoded calcium indicator (GCaMP6f) that is selectively expressed in cortical astrocytes to monitor the Ca 2+ transients with two-photon microscopy 2 ( Fig.1A-B, Figure S1). Two weeks after virus microinjections, GCaMP6f-expressing astrocytes can be seen under two-photon microscopy, similarly as our previous reports 31 . On the day of the experiment, a cranial window was prepared under 1.5% isoflurane anesthesia. The animals totally recovered after 30 min prior to in vivo whole-cell recordings of single neurons and monitoring astrocytes in layer 2/3 of barrel cortex ( Fig. 1A) 9,28,31 . The sleep states were defined as periods in which the animals closed eyes, but spectral analysis of electroencephalography (EEG) recordings showed that most of its power resides in slow waves of 0.5~4Hz 32 (Fig. 1C); concurrently with relatively low amplitude in electromyography (EMG) recordings in the neck.
In both sleep and waking mice, we were able to identify some spontaneous Ca 2+ signals in fine processes. Representative fine processes were randomly chosen to analyze the fluorescence changes (Fig. 1D). As expected, low frequency of whisker simulation (5 Hz) significantly increased astrocytic Ca 2+ transient in the astrocytic fine processes at both sleep and waking states.
However, the whisker stimulation induced Ca 2+ transients during sleep state are much smaller compared with that during the waking state (peak fluorescent peak increased 23.2 ± 14.4% for whisker stimulation at sleep state, compared with 73.3 ± 11.7% for at arousal states; p < 0.01, ttest, n = 24 processes in 6 animals, Fig.1D-F; Figure S2). Awake states were characterized as eyes opening, spontaneously whisker movement, and accompanied with typical electromyography (EMG) recordings (Fig.1C). Whisker stimulation induced much larger Ca 2+ transients at arousal state than those during sleep state, with a fast latency (3.8 ± 2.1 s compared with 7.5 ± 1.6 s at sleep state, n = 24 processes in 6 animals, Fig. 1G), a fast rise of fluorescence increase (5.8 ± 2.4 %/s compared with 22.3 ± 4.8 %/s at sleep state, p < 0.01, paired t-test, n = 24 processes in 6 animals, Fig. 1H). However, the decay was also faster for the astrocytic Ca 2+ signals at arousal state (4.3 ± 0.8 %/s compared with 0.8 ± 0.7 %/s at sleep state, p < 0.01, paired t-test, n = 24 processes in 6 animals, Fig. 1I); while the duration were similar (with 12.3 ± 3.5 s compared with 18.6 ± 4.5 s for sleep, n = 24 processes in 5 animals; paired t-test, p = 0.56).

The small Ca 2+ transients blocked sensory transmission
To monitor the effects of the small Ca 2+ transients on neuronal activity, we did in vivo whole-cell recordings for the neurons in layer II -III. During sleep, all neurons stereotypically oscillated between two intrinsic stable membrane potentials (1.23 ± 0.25 Hz, -75.4 ± 2.7 mV and -54.7 ± 3.2 mV, n = 7 cells in 7 animals, one cell in each animal, Figure S3). However, contrary to previous reports that the thalamus can block the sensory transmission to the cortex, whisker stimulation really elicited EPSPs during sleep, which can be recorded in both whole-cell recording (averaged at 13.4 ± 2.1 mV mV for the first ten EPSPs, n = 8 cells, Fig. 3A), and LFP recordings (averaged at 0.8 ± 0.4 mV mV for the first ten EPSPs, n = 8 cells, Fig. 3B). The whole-cell EPSPs happened at 29.8 ± 1.6 ms after air puffing (n = 8 cells, one cell in each animal), while the LFP EPSPs occurred with a latency of about 19.3 ± 1.7 ms after whisker stimulation (Fig.3C). However, the LFP EPSPs terminated much earlier than Vm EPSP (52.38 ± 6.2 ms, compared with 186.4 ± 17.4 ms for whole-cell recordings, n = 8 animals, t-test, p < 0.01; Fig.3C). Of note, the latency of the evoked response might be extremely long, which might be due the time of air puffing. Then we used ATP and UTP to induce the Ca 2+ , and found that the UTP induced Ca 2+ transient in the astrocytes had little effects on the slow oscillation (if not removed the up-down state oscillation), but blocked the sensory transmission. After the Ca 2+ transients, the whisker stimulation induced EPSPs decreased (Fig.3D). We also screened whisker stimulation induced Ca 2+ transients, by first apply a batch of whisker simulation (10 s, 5 Hz), and after 5 seconds, apply another batch of whisker stimulation, and the results showed that the second batch of stimulation got much smaller EPSPs (13.8 mV ± 2.3 mV as before, compared with 6.8 mV ± 1.8 mV after, paired t-test, p < 0.01, n = 5 animals). Similarly, the EPSPs recorded in LFP also decreased from 0.82 mV ± 0.18 mV to 0.38 mV ± 0.14 mV. In addition, the EPSP decrease can be reversed by blocker SN-6 (50 µM) and SEA0400 (50 µM), and BAPTAam (10 µM) during the sleep states (One-way ANOVA, p < 0.01, n = 5-7 animals, Fig. 3E). To ensure the astroglial origin of the EPSP enhancement, agonist-induced Ca 2+ signals were studied in MrgA1 + mice 33

Big astrocytic Ca 2+ transients at arousal states enhanced sensory transmission
Then we investigated the mechanisms of astrocytic Ca 2+ signals at awake states. Whisker stimulation induced much larger Ca 2+ transients at fine processes during arousal state than those during sleep state, the fluorescence increase was 73.3 ± 11.7%, with a duration to be 18.6 ± 4.5 s (n = 24 cells in 5 animals, Fig. 4A-B.). The characteristics of astroglial Ca 2+ transients was tested in the presence of mGluR antagonist MPEP (50 µM), and it is found that the Ca 2+ transients, ΔF/F0 decreased to 17.7 ± 9.2 % during waking states (21 cells, in 5 animals, Fig. 4B). This suggested that the astroglial Ca 2+ transients at waking states are due to glutamate release. In addition, when the mice with genetic deletion of IP3 receptor (Type 2), whisker stimulation induced Ca 2+ transients decreased significantly to 25.5 ± 5.4 % (n = 19 cells in 4 animals, Fig.   4B). Contrary to that of sleep, NCX blocker SN-6 (50 µM) and SEA-0400 (50 µM) has little effects on whisker stimulation induced Ca 2+ transient, with an averaged peak fluorescence of 58.6 ± 5.1 % (36 processes, in 5 animals, Fig. 4B). In addition, CRAC channel blocker CM-4620 has little effects either. These data suggest that the big Ca 2+ transient is due to IP3 dependent Ca 2+ release from endoplasmic reticulum (ER).

Adrenergic activity is a prerequisite for big Ca 2+ waves
As the whisker stimulation induced smaller Ca 2+ signaling during sleep state and bigger Ca 2+ signaling during arousal waking state, and their functions seem to make the sleep brain even quieter, the waking brain even more aroused. We next try to see if they can switch from one to the other, like previous report, which suggested that the astrocytic Ca 2+ needs the norepinephrine (NE) in locus cereus (LC) to be primed 28 . First we stimulated LC during sleep, and then applied a batch of whisker stimulation, and compared the whisker stimulation induced Ca 2+ transients before and after LC stimulation. Whisker stimulation during sleep states induced small increases in Ca 2+ transients; and after LC stimulation, the animals really woke up, and then whisker stimulation induced much larger Ca 2+ transients (fluorescence transient was 27.2 ± 4.3% for whisker stimulation before LC stimulation; after LC stimulation, Ca 2+ transients increased by 68.8 ± 6.1%, n = 27 cell processes in 5 animals, ** one-way ANOVA, p < 0.01) (Fig. 5A-B). To further prove the LC stimulation was really through NE, we injected NE (100 µM) to the barrel cortex, and found that whisker stimulation induced Ca 2+ signaling really increased from 23.6 ± 3.8% to by 61.8 ± 4.1%, n = 29 cell processes in 5 animals, ** one-way ANOVA, p < 0.01). We then used different adrenergic antagonists to the Barrel cortex, such as α1-AR antagonist terazosin (100 µM). Terazosin significantly blocked whisker timulation induced Ca 2+ signalling (peak of ΔF/F0 in the presence of the drug was 31.2 ± 4.9%, p < 0.01, t-test, n = 31 cells in 4 animals). Conversely, α2-AR antagonist metoprolol (10 µM) was not significantly effective (ΔF/F0 peaked at 52.7 ± 7.6%, n = 28 cells in 4 animals) at LC induced Ca 2+ signaling.
We also compared whisker stimulation evoked EPSPs before or after LC stimulation, and found that LC significantly increased the amplitude of whisker stimulation induced EPSPs (from 0.51 ± 0.12 mV for the first batch of whisker stimulation, to 1.1 ± 0.19 mV for the second batch of whisker stimulation, n = 5, paired t-test, p < 0.01 (Fig. 5B). To make sure the enhancement of EPSPs was due to Ca 2+ transients induced by first batch whisker stimulation, we applied BAPTA/AM (10 µM), which was majorly taken up by astrocytes, and has been often used to block astrocytic Ca 2+ transients in many previous reports 33 . The results showed that ΔF/F0 decreased to 12.7 ± 1.2 % in BAPTA/AM loaded cells after LC stimulation, and the EPSPs were also decreased (from 0.78 ± 1.2 mV to 0.3 ± 0.2 mV, **p < 0.01, One-way ANOVA, 5 animals, Fig. 5C). In addition, α1-AR antagonist terazosin (100 µM) significantly reduced EPSPs (before 0.91 ± 0.12 mV to 0.41 ± 0.19 mV, n = 5 animals. Fig. 5D-E). Reversely, we also found that when the animals turned from arousal state to sleep states, the EPSPs got smaller (Fig. S4).

Discussion
Increasing body of evidence has proved that astrocytes are major players in the modulation of In addition, astrocyte has been suggested to be involved in attention control 41 . As far as we know, these two kinds of astrocytic Ca 2+ signaling during sleep state and the aroused waking states have not been reported before. Here we further found that the LC stimulation acts switch between the small Ca 2+ signaling at sleep state to the higher Ca 2+ signaling, which might be the mechanism for LC/NE modulating brain arousal state 42 .

Small Ca 2+ transient facilitates sleep like slow oscillation during sleep state
During sleep, whisker stimulation induced K + release from neurons, will be taken up by astrocytes, K + influx would in turn induced Ca 2+ influx; this kind of Ca 2+ influx or ATP agonist induced Ca 2+ release can open Ca 2+ release activated Ca 2+ channels (CRAC) to evoke Ca 2+ transients (Figure 6). Astrocytic processes contains many microdomains that contains the NCX and Na + pumps with high ouabain affinity α2, and the increase of Ca 2+ would in turn increase NCX activity, thus increase sytosolic Na + and Ca 2+ concentrations; and increased sytosolic Na + in turn activate Na + pump to increase K + take-up by astrocytes 19 . Thus astroglial Ca 2+ transients would decrease extracellular K + , to hyperpolarize neurons and block sensory transmission 20 . In all, astrocytes block the sensory transmission possibly by the membrane potential hyperpolarization during sleep.

Large Ca 2+ wave enhances sensory transmissions
Here we might be the first to report that the astrocytes in the barrel cortex exhibit two types of Ca 2+ events. These two types of events appear to have different pharmacology/sources and may play a different role in facilitating the efficacy of synaptic transmission during different behavior states. During last three decades, it has been found that astroglial cells regulate synaptic connectivity through multiple mechanisms [40][41][42][43][44] , including the concept of "active" astrocytes capable of releasing 'gliotransmitters' 45-47 . Overwhelming evidence for such physiological properties have confirmed the effects of gliotransmitters on synaptic activities 48 . Therefore, the reported results about function of Ca 2+ signaling in this study are possibly due to glutamate release to regulate synaptic connectivity. Even though some studies about astrocytic Ca 2+ suggests that astrocyte calcium transients are not involved in synaptic transmissions in hippocampus 7 . The "third wave" of astrocytic calcium signaling research in astrocytic fine processes has now been surged up, and focused on the calcium transients occurring in fine astrocyte processes not resolved in earlier studies 44 . Here our result report that the large Ca 2+ wave is really a prerequisite for reliable sensory transmission at waking states.

Adrenergic activation acts as a switch to evoke large Ca 2+ wave
Previous studies suggested that astrocyte responds with robust calcium elevations during arousal/startle due to the release of noradrenaline from noradrenergic projections via a mechanism likely involving α1 adrenoceptors on astrocytes 17 . Here we also report that astrocytic calcium transients are normally quiet 27 and very hard to be activated without synergetic activation by NE/LC system together whisker stimulation. NE, an important neurotransmitter in the circadian rhythm, can prime astrocytic Ca 2+ signaling, which in turn induces many astrocytic activities such as gliotransmitter releases (Figure 6). NE acts on astrocytes through αand βadrenergic receptors (α/β−ARs), with the former being linked to InsP3 production and intracellular Ca 2+ release 46 . Our data, together with our previous data also showed that it is through αadrenergic receptors (α/β−ARs) that the astrocytic Ca 2+ wave is primed [49][50][51] . One of the major functions of NE induced astrocytic Ca 2+ wave is increasing extracellular K + 52 . After LC stimulation or NE application, the extracellular K + can increase immediately approximately 1 mM 20 . The increased K + can depolarize the neurons to let them leave the slow oscillation and be ready to fire action potentials at EPSPs. The increased extracellular K + can also prime astrocytic Ca 2+ to increase its function at clearance and recycle of the neurotransmitters, help the synaptic transmission last longer. In all, astrocytic calcium signaling is a prerequisite for reliable sensory transmission, especially during waking state when the neurons need more work for astrocytes 53 .
In addition, the glymphatic system is a glial-dependent perivascular network that plays a pseudolymphatic role in the brain. The perivascular spaces form a complex brain-fluid transport system that supports fast exchange with interstitial fluid and clearance of waste products from the brain from the intricate environment of the neuropil. Sleep is necessary for the normal function of the glymphatic system; while norepinephrine release blocks the glymphatic system.
Norepinephrine release decreases during sleep, resulting in an expansion of the interstitial space and a subsequent potentiation of glymphatic-fluid transport. A recent study suggested that the astrocytic Ca 2+ elated with the glymphatic systems 3 , which suggested that neural slow waves are followed by hemodynamic oscillations, which in turn are coupled to CSF flow. These results demonstrate that the sleeping brain exhibits waves of CSF flow on a macroscopic scale, and these CSF dynamics are interlinked with neural and hemodynamic rhythms 3 .

Materials and Methods
Animal preparation for awake in vivo recordings: Adult (10 weeks old) C57Bl/6 wild type mice were used (both male and female, from Charles River Laboratories). The preparation for mouse experiments was modified from published protocols 48,49 . Briefly, mice were anesthetized using isoflurane (1.5% mixed with 1-2 L/min O2), head restrained with a custom-made miniframe and habituated to the restraint over one week in multiple session, with a total training duration of 3-4 hours. A 1.5 mm craniotomy was then opened over the somatosensory cortex (1.5 mm in diameter, 3 mm lateral and 1.5 mm posterior to the Bregma), the dura was carefully removed, and the mice were allowed 60 min recovery prior to conducting the experiments. The craniotomy procedure lasted < 20 min to minimize anesthesia exposure on the recording day. Animals were then head-strained, placed in a behavioral tube to minimize movement and relocated to the imaging room, which was kept dark and sleep. Body temperature was maintained with a heating pad. For cortical drug surface application, artificial cerebrospinal fluid (aCSF) was perfused across the cortex of awake mice at a rate of 2 mL/min, into a custom-made well with ~200 μL volume, through tubing with ~100 μL volume, meaning the entire volume bathing the brain was exchanged every ~9 s. The aCSF solution contained 126 mM NaCl, 2.5 mM KCl, 1.25 mM NaH2PO4, 2 mM MgCl2, 2 mM CaCl2, 10 mM glucose, and 26 mM NaHCO3, pH 7.4. For imaging, calcium indicator rhod-2 AM (Invitrogen) was loaded onto exposed cortex for 30-40 min before applying agarose (1.5%, type III-A, Sigma) and a coverslip 50 . All animals were trained several times a day, for three days before the experiments. All procedures followed National Institute of Health guidelines and were approved by the Institution of Animal Care and Use Committee. The animals were anesthetized with 1.5% isoflurane during the surgery, and then after being mounted on the equipment, the isoflurane was stopped and the animal was left to run freely on the wheel during sleep/wake states in the dark room, except that the heads were held tightly on a frame (detailed methods refer to previous publications, e.g. 9 ). Surgical procedure for virus injection: pAAV5-GfaABC1D-cyto-GCaMP6f-SV40 (Cat# AV-5-52925, UPenn vector core), which is a genetically encoded Ca 2+ indicator driven by the astrocyte specific GfaABC1D promoter, was injected in Barrel cortex. The mice were anesthetized with 1-2% isoflurane with oxygen supply and were placed in a stereotactic head frame with a heating pad underneath. A small vertical incision was made on the skin and the craniotomy (0.5 mm × 0.5 mm) was performed with a drill. The virus was injected at a volume of 500 nl/site without dilution, 3 injection sites were utilized. A glass micropipette with a tip of 10 µm was used for injection with microinjection control.
In vivo two-photon imaging and stimulation ： A custom-built microscope attached to a Tsunami/Millenium laser, Spectra Physics, Mountain View, CA) and scan box (FV300 Fluoview Software, Olympus, Center Valley, PA) was used for 2-photon imaging through a 203 objective (0.9 NA, Olympus). Excitation wavelength was in the range of 800--820 nm. Emission wavelengths were split to detect fluo-4 and AlexaFluor 594 signals as previously described 23 . Images of astrocytic Ca 2+ signaling were recorded every 2--3 s, which was sufficient to capture evoked responses while limiting laser induced photo damage at a laser power of <30 mW. Prior to whisker stimulation experiments, anesthesia 0.5 mg/kg D-tubocurarine was injected to prevent small reflex movements that could distort imaging. Direct LC stimulation was applied using a bipolar concentric electrode. Stimulation consisted of a single train of 20--100 pulses (100 Hz, 50 µA, 0.5 ms square pulses).
In vivo whole-cell recording: Recordings were obtained from layer II barrel cortex using glass microelectrodes. LFP signals were externally filtered at 6 Hz (Filter Butterworth Model by Encore, Axopatch 200B by Axon Instruments), bandpass filtered at 1-100 Hz and digitized (Digidata 1440A by Axon Instruments). Recordings were analyzed offline using pClamp 10.2. Whole-cell recordings were performed with blind patching, by watching the pipette resistance. Patch electrodes were fabricated from filament thin-wall glass (World Precision Instruments) on a vertical puller; the resistance of the pipette was about 6 to 9 megohms with intracellular pipette solution added. The pipette solution contained 140 mM K-gluconate, 5 mM Na-phosphocreatine, 2 mM MgCl2, 10 mM Hepes, 4 mM Mg-ATP, and 0.3 mM Na-GTP (pH adjusted to 7.2 with KOH). The junction potential between the patch pipette and the bath solution was zeroed before forming a gigaseal. Patches with seal resistances of less than 1 gigohm were rejected. Data were low pass-filtered at 2 kHz and digitized at 10 kHz with a Digidata 1440 interface controlled by pClamp Software (Molecular Devices). Whisker stimulation was delivered using a picospritzer III (Parken Instrumentation) and Master 8 (A.M.P.I.). The amplifier bandwidths were normally 0.5 Hz to 100 Hz. EEG recording was digitized at 100 Hz and then subjected to spectral analysis using a complex demodulation procedure.
In vivo intracellular recording: In vivo intracellular recordings were obtained from layer II barrel cortex using glass microelectrodes. The surgery was done the same as in vivo whole cell recordings. The electrode was made from a thick wall glass (B150F-4, World Precision Instruments). The resistance of the electrodes was50-150 MΩ, after filled with 1.0 M potassium acetate. Intracellular recording was done with the same equipment as whole cell recordings, except that the manipulation of the pipette was done with the advancement of 1-2 µm step with the fastest movement mode. After penetration, the series resistance was adjusted, and the currentclamp was used to measure the membrane potentials of the astrocytes.
EEG and EMG recording Acquired EEG/EMG signals were amplified ((Filter Butterworth Model by Encore, Axopatch 200B by Axon Instruments), at a sampling frequency of 1K Hz. The EEG signal was filtered with high-pass: 0.5 Hz, low pass: 30 Hz, and EMG signal high-pass filtered at 10Hz. Wakefulness was subdivided into sleepfulness (DW) or arousal wakefulness (AW) using EMG peak-to-peak amplitude of all wake epochs across the 12-h recording. QW was defined as 33rd percentile or less and AW 66th percentile or higher of all wake EMG peakto-peak amplitude values. Concurrently with EEG recordings, spectral analysis of electroencephalography (EEG) recordings showed that most of its power resides in 4~6Hz, which were interrupted intermittently with slow waves of 0.5~4Hz.
Statistical analysis: All analyses were performed using SPSS 19 software (IBM) and all tests were two-tailed where significance was achieved at α = 0.01 level. For independent samples, a ttest (≤ 2 variables) or one-way ANOVA (> 2 variables) was used; for paired samples, a paired t test was used.   Typical traces show the Ca 2+ transients and also changes of extracellular K + . C. Statistical data show the comparison the whisker stimulation induced changes of Ca 2+ transients with agonists, and also in IP3R2 knockout mice, and also with different manipulation of blockers (p < 0.01, one-way ANOVA, n = 4 -6 animals). D. A schema shows the intrinsic ion channels involved in the oscillation. Whisker stimulation induced K + release from neurons, will be taken up by astrocytes via Kir4.1, K + influx induced Na + influx, which in turn induced Ca 2+     Upper traces (blue trace is from sleeping state, while the red trace is after LC stimulation) are relative changes of Ca 2+ (ΔF/F0) in an astrocytic process located close to the recording electrode. Lower traces are LFP recorded by the recording electrode. D. Analysis shows comparisons of the relative changes of Ca 2+ (ΔF/F0) in astrocytic processes at whisker stimulation (**, p < 0.01, one-way ANOVA, n = 4-7 animals). E. Analysis shows comparisons of the relative changes of LFP EPSPs at whisker stimulation (**,p < 0.01, one-way ANOVA, n = 4-7 animals). The x-axis labels indicate the type of stimulation used. Additional abbreviations used are as following: TER, α1-antagonist terazosin; METO, β-antagonist metoprolol; ** p < 0.01. Figure 6. A model shows the mechanisms of Ca 2+ transients during sleep and waking states. Left. During sleep, whisker stimulation induced K + release from neurons, which will be taken up by astrocytes, K + influx induced Ca 2+ influx, or ATP agonist induced Ca 2+ release can activate Ca 2+ release activated Ca 2+ channels (CRAC) open to get much more Ca 2+ influx inside the processes. The Ca 2+ increase would in turn make feedback to increase intracellular Na + , which would activate Na pump to take up K + , and make the neurons to hyperpolarization. Right. During waking states, NE activates Gq receptors which primes the astrocytes, and synergistically works with mGLu to activate Gq receptors to release even more IP3, and induce Ca 2+ release from ER, and induced gliotransmitters release (such as ATP, or glutamate, or glutamine).   The lower trace is the extension of the recordings in the square, which shows that the down-state and up-state can be reversed. B. The I-V curves drawn from the traces in A, show the possible ion channels involved in down-state and up-state, which might be leaking K + currents and persistent Na + /Ca 2+ currents respectively. C. A model shows the intrinsic ion channels involved in the oscillation. The leaking K + currents are outward currents, and the persistent Na + /Ca 2+ currents are inward, their balance constitutes the two stable states: up-states and down-states.