Norepinephrine links astrocytic activity to regulation of cortical state

Reitman, Michael E.; Tse, Vincent; Mi, Xuelong; Willoughby, Drew D.; Peinado, Alba; Aivazidis, Alexander; Myagmar, Bat-Erdene; Simpson, Paul C.; Bayraktar, Omer A.; Yu, Guoqiang; Poskanzer, Kira E.

doi:10.1038/s41593-023-01284-w

Download PDF

Article
Open access
Published: 30 March 2023

Norepinephrine links astrocytic activity to regulation of cortical state

Nature Neuroscience volume 26, pages 579–593 (2023)Cite this article

16k Accesses
19 Citations
202 Altmetric
Metrics details

Subjects

This article has been updated

Abstract

Cortical state, defined by population-level neuronal activity patterns, determines sensory perception. While arousal-associated neuromodulators—including norepinephrine (NE)—reduce cortical synchrony, how the cortex resynchronizes remains unknown. Furthermore, general mechanisms regulating cortical synchrony in the wake state are poorly understood. Using in vivo imaging and electrophysiology in mouse visual cortex, we describe a critical role for cortical astrocytes in circuit resynchronization. We characterize astrocytes’ calcium responses to changes in behavioral arousal and NE, and show that astrocytes signal when arousal-driven neuronal activity is reduced and bi-hemispheric cortical synchrony is increased. Using in vivo pharmacology, we uncover a paradoxical, synchronizing response to Adra1a receptor stimulation. We reconcile these results by demonstrating that astrocyte-specific deletion of Adra1a enhances arousal-driven neuronal activity, while impairing arousal-related cortical synchrony. Our findings demonstrate that astrocytic NE signaling acts as a distinct neuromodulatory pathway, regulating cortical state and linking arousal-associated desynchrony to cortical circuit resynchronization.

Distinct astrocytic modulatory roles in sensory transmission during sleep, wakefulness, and arousal states in freely moving mice

Article Open access 17 April 2023

Distinct temporal integration of noradrenaline signaling by astrocytic second messengers during vigilance

Article Open access 24 January 2020

Centripetal integration of past events in hippocampal astrocytes regulated by locus coeruleus

Article Open access 03 April 2024

Main

Patterns of neural activity in the awake cortex are variable¹, ranging from states of highly synchronized neuronal activity during periods of low arousal, to states of desynchronized activity during high-arousal periods such as whisking² and running³. These cortical states can be identified by electrophysiological activity: synchronized cortical states display greater low-frequency (LF) oscillations and reduced high-frequency (HF) oscillations than desynchronized states⁴. While HF oscillations are important for processing of incoming information⁵, LF oscillations during the wake state are less well characterized. Generally described within a 2–10 Hz range^2,6,7, wakeful LF power differs substantively from cortical activity patterns during sleep², indicating a unique role for synchronized cortical states in awake animals. One role of waking cortical synchrony may be to modulate sensory responses to external stimuli. Increased LF power is associated with reduced neuronal gain⁶ and broader tuning⁸ in sensory cortex, as well as diminished behavioral performance on sensory perception tasks^3,7,8.

The mechanisms by which cortical state is regulated, and how the cortex becomes sensitized to external stimuli, are well studied. One central mechanism is signaling by norepinephrine (NE), which regulates a primary hallmark of behavioral arousal—pupil diameter⁹—and alters the firing properties of cortical neurons leading to desynchronized cortical activity^10,11. However, while the desynchronizing effects of NE are well known, mechanisms that increase waking synchrony are less clear. A full understanding of mechanisms that regulate awake cortical synchrony will be necessary to explain fluctuations in perception and behavior.

Neurons are not the only NE-responsive cell types in the cortex^12,13. Astrocytes, a non-neuronal cell type abundant throughout the cortex, respond to NE with robust calcium (Ca²⁺) signaling^{14,15,16,17,18,19}. However, astrocyte Ca²⁺ has also been linked to increased cortical synchrony when NE signaling is low, including during sleep²⁰ and under anesthesia²¹. In this Article, we aimed to resolve the disparity between an astrocytic role in sleep generation and their activation by NE and desynchronizing stimuli such as movement. We hypothesized that NE-specific astrocytic signaling might act as a mechanism that regulates NE-driven cortical desynchrony and restores cortical synchrony following changes in arousal.

Here we use in vivo, two-photon (2P) Ca²⁺ imaging to show that astrocytes in mouse visual cortex respond proportionally and with temporal specificity to changes in arousal and NE, positioning astrocytes as a local feedback mechanism to the effects of arousal. In agreement with this hypothesis, NE-driven astrocyte Ca²⁺ signaling occurs alongside reductions in arousal-associated neuronal activity. Using in vivo local field potential (LFP) recordings, we further show that arousal-driven astrocyte Ca²⁺ activity occurs at transitions from cortical desynchrony to synchrony, and that this relationship is dependent on NE signaling. Pharmacological stimulation of Adra1a receptors counterintuitively increases wakeful cortical synchrony, and astrocyte-specific removal of Adra1a enhances total and arousal-driven neuronal activity, while impairing arousal-related cortical synchrony. Our results directly link astrocytic NE receptors to cortical state regulation. We thus identify NE signaling to astrocytes as a new circuit mechanism by which astrocytes act as sensors of NE changes and synchronize the cortex in response to arousal.

Results

Astrocyte Ca²⁺ correlates with increases in pupil diameter

To first determine whether astrocyte Ca²⁺ activity is dynamically modulated by arousal, we carried out in vivo, 2P Ca²⁺ imaging in visual cortex of awake, head-fixed mice while simultaneously recording pupil diameter and running speed (Fig. 1a). We expressed the Ca²⁺ indicator GCaMP in cortical astrocytes under the GfaABC₁D promoter and used the Astrocyte Quantitative Analysis (AQuA) toolkit²² to accurately capture dynamic fluorescent astrocyte signals, even from spatially overlapping events (Extended Data Fig. 1a,b).

**Fig. 1: Changes in arousal shape astrocyte Ca²⁺ activity independent of movement.**

We extracted fluorescence traces from all AQuA-detected Ca²⁺ events and found that movement was accompanied by large increases in astrocyte Ca²⁺—as well as increases in arousal indicated by pupil dilation—consistent with previous reports^{6,14,16,23,24} (Fig. 1b, left). Even during stationary periods, astrocyte Ca²⁺ activity was accompanied by concurrent and proportional fluctuations in arousal (Fig. 1b, right). Population astrocyte Ca²⁺ activity was better correlated with pupil diameter (0.51 ± 0.09) than running speed (0.21 ± 0.07) when comparing across mice (Fig. 1c, top, n = 6 mice) or using hierarchical bootstrapping (HB, Fig. 1c, bottom). In addition, individual astrocyte Ca²⁺ events correlated better with arousal compared with either speed- or time-shuffled data (Extended Data Fig. 1c, left). Most Ca²⁺ events (9,597/11,674; 82.2%) exhibited a maximal cross-correlation with pupil diameter within 10 s and a short (0.32 ± 0.01 s standard error of the mean (s.e.m)) lag, consistent with arousal driving astrocyte activity (Extended Data Fig. 1c, right).

Our analysis indicates that arousal contributes to astrocyte Ca²⁺ beyond its association with movement. However, cortical activity reflects both absolute pupil diameter²³ and relative changes in pupil diameter²⁴. To investigate whether astrocytes sense absolute levels of arousal, we binned the pupil diameter into deciles (Fig. 1d, left), and calculated the average Ca²⁺ fluorescence in each. Astrocyte Ca²⁺ dynamically varied with pupil diameter only when the pupil was ~40–80% of maximum diameter (Fig. 1d, right, and Supplementary Table 1). This range matched the overlap in pupil diameter found between stationary and movement periods (Extended Data Fig. 1d), indicating this relationship was a function of movement, and suggesting that absolute pupil diameter did not adequately explain the relationship between arousal and astrocyte Ca²⁺.

To test whether relative changes in pupil diameter were linked to astrocyte Ca²⁺ responses, we calculated an event-triggered average relative to the start of movement. As expected, we found that pupil diameter and—with a short delay—astrocyte Ca²⁺ increased around movement onset (Fig. 1e). We then asked which behavioral variables best explained movement-evoked astrocyte Ca²⁺ activity (Fig. 1f–h). We found that speed was a poor predictor of the maximal astrocyte Ca²⁺ fluorescence (Fig. 1f, left) as previously reported¹⁶, as was the relative change in speed (Fig. 1f, right). In addition, absolute pupil diameter following movement had a weak correlation with astrocyte Ca² fluorescence, indicating that, even during behavioral state changes, absolute pupil diameter is a poor predictor of astrocyte responses (Fig. 1g, left). In contrast, the relative change in pupil diameter explained a substantial portion of the astrocyte Ca²⁺ response to movement, indicating that astrocytes are specifically sensitive to relative changes in arousal (Fig. 1g, right).

We further dissociated the effects of movement and arousal on astrocytes by separating behavioral state periods (Extended Data Fig. 1e). The relationship between astrocyte Ca²⁺ and increases in pupil diameter persisted during stationary periods (Fig. 1i–j), even though pupil diameter changes were smaller (Extended Data Fig. 1f). These results demonstrate that astrocytes are sensitive to smaller changes in arousal than previously recognized.

We also wondered whether changes in arousal could explain the timing of astrocyte Ca²⁺ activity. We noticed that astrocyte Ca²⁺ peaked with pupil diameter around movement offset (Fig. 1k) and the time to maximum pupil diameter was strongly correlated with the time to peak astrocyte activity (Fig. 1l). Both were also dependent on the duration of the movement bout (Extended Data Fig. 1g). When examining all astrocyte Ca²⁺ events, onsets occurred more often during dilation and offsets during constriction (Fig. 1m). These findings indicate that changes in arousal shape the timing and level of astrocyte Ca²⁺ activity.

Phasic increases in NE precede astrocyte Ca²⁺

NE is a key driver of both changes in pupil diameter and brain activity²⁵. While recent work suggests that astrocytes preferentially respond via Ca²⁺ to “multi-peaked” NE axonal activity¹⁵, the impact of behavioral state has not yet been considered. To determine the relationship between NE and astrocyte Ca²⁺ activity, we simultaneously expressed a fluorescent sensor of NE, GRAB_NE²⁶, in cortical neurons under the h-syn promoter and the Ca²⁺ indicator jRGECO1b in cortical astrocytes under the GfaABC₁D promoter (Fig. 2a and Supplementary Video 1). Examining the raw GRAB_NE fluorescence, we saw reductions in fluorescence that matched changes in background vasculature, independent of indicator expression (Extended Data Fig. 2a). Our 2P imaging paradigm was not compatible with techniques to directly measure hemodynamic effects such as reflectance imaging and blind-source separation²⁷. Therefore, we developed a method to approximate and compensate for hemodynamic signals in 2P imaging (Extended Data Fig. 2b and Methods). To further limit hemodynamic contamination, we took the mean of the corrected GRAB_NE signal, excluding highly contaminated or artifactual pixels (Extended Data Fig. 2c). As our methodology only approximates hemodynamic artifacts and does not account for possible hemodynamics effects from excitation or out-of-plane GRAB_NE fluorescence, we next sought to confirm the accuracy of our methodology. We found our corrected GRAB_NE signal showed little correlation with vasculature or hemodynamically contaminated regions (r < 0.2) but correlated well with regions that had active GRAB_NE signal (Extended Data Fig. 2d). To further validate our methodology in an unbiased manner, we correlated individual pixels with pupil diameter (Extended Data Fig. 2e) and found the corrected GRAB_NE signal linearly correlated with pupil diameter-related pixels suggesting our method reflects GRAB_NE signal that occurs from pupil-related increases in NE while reducing hemodynamic effects (Extended Data Fig. 2f).

**Fig. 2: Astrocytes are sensitive to a range of NE increases.**

Using our corrected GRAB_NE signal, we found that GRAB_NE fluorescence matched changes in pupil diameter and astrocyte Ca²⁺ activity, although the GRAB_NE showed a slow decay, similar to previous reports¹⁵ (Fig. 2b, left). This relationship persisted even during stationary periods (Fig. 2b, right). The slow GRAB_NE decay was reflected in the power spectrum, which showed an inverse relationship with frequency (Fig. 2c). Slow (0.01–0.03 Hz) GRAB_NE fluctuations had more power than the empirically fit 1/f relationship, suggesting long-lasting fluctuations in “tonic” NE predominate the GRAB_NE signal.

To confirm the link between GRAB_NE and arousal, we examined the cross-correlation between pupil diameter and GRAB_NE. We found a positive correlation between GRAB_NE and spontaneous changes in pupil diameter (r = 0.13, Fig. 2d), similar to that previously described for cortical NE axons²⁸. However, the maximum cross-correlation was broad and continued for at least 20 s after the pupil diameter (Fig. 2d, left). Further, unlike what has been reported for NE axons²⁸, GRAB_NE showed poor cross-correlation (r = 0.04) with the derivative of pupil diameter (Fig. 2d right, and 2e). These results suggest that cortical GRAB_NE dynamics best report arousal level rather than changes in arousal, and may substantively differ from the activity pattern of the underlying NE axons.

We next quantified the relationship between GRAB_NE and astrocyte Ca²⁺ activity. We found a positive, broad correlation (r = 0.17) between the two (Fig. 2f, left), with GRAB_NE preceding astrocyte Ca²⁺ (−5.7 ± 3.1 s delay, Fig. 2f, right) and GRAB_NE showing a similar link to astrocyte Ca²⁺ as arousal (Fig. 2g). Although slow NE fluctuations predominated our GRAB_NE signal (Fig. 2c), we also noticed phasic increases in NE (Fig. 2h, left, triangles). We segregated phasic NE peaks by amplitude (Fig. 2h, left, triangles) and saw that larger phasic GRAB_NE activity was also longer lasting (Fig. 2h, right, and Supplementary Table 2). We then examined how astrocyte Ca²⁺ changed following phasic NE activity. We saw that large (≥2 standard deviations (s.d.)) phasic changes in NE preceded prolonged increases in astrocyte Ca²⁺ (Fig. 2i–j, left, and Supplementary Table 3, left), while small increases in GRAB_NE co-occurred with proportional increases in astrocyte Ca²⁺ (Fig. 2i, right). These increases were observed even with smaller (1 s.d.) changes in NE (Fig. 2j, right, and Supplementary Table 3, right). These results suggest astrocytes dynamically respond to phasic changes in NE, in both the duration and amplitude of their Ca²⁺ signaling. We also calculated the event-triggered average GRAB_NE signal relative to astrocyte Ca²⁺ events and confirmed that GRAB_NE increased before astrocyte Ca²⁺, and peaked shortly after astrocyte Ca²⁺ signaling had begun (Fig. 2k). This relationship was true for both astrocyte Ca²⁺ events during movement (Fig. 2k, left) and stationary periods (Fig. 2k, right). In sum, these results indicate that astrocytes are sensitive to changes in NE across behavioral states.

To corroborate these findings, we performed freely moving, dual-color fiber photometry recordings of GRAB_NE and astrocyte Ca²⁺ (Extended Data Fig. 3a) and saw similar GRAB_NE dynamics (Extended Data Fig. 3b,c). To confirm the relationship between NE and astrocyte Ca²⁺, we used tail lifts to evoke startle responses¹⁴, and saw expected increases in both GRAB_NE and astrocyte Ca²⁺ (Extended Data Fig. 3d). The GRAB_NE signal began increasing before astrocyte Ca²⁺ and persisted after the startle response (Extended Data Fig. 3e). These results further support the hypothesis that astrocytes respond to phasic increases in NE.

Astrocyte Ca²⁺ may reduce arousal-driven neuronal activity

We next wondered how arousal-mediated changes in nearby neurons relate to astrocyte Ca²⁺ activity, as arousal also strongly modulates neuronal activity^6,24,29. To answer this, we expressed hSyn-GCaMP6f in neurons and GfaABC₁D-jRGECO1b in astrocytes to record the Ca²⁺ activity of both cellular populations simultaneously (Fig. 3a,b and Supplementary Video 2).

**Fig. 3: Astrocyte Ca²⁺ is positioned to reduce effects of arousal on population-level neuronal activity.**

We found that astrocyte and neuronal Ca²⁺ fluctuated together along with movement and pupil diameter (Fig. 3c). The average fluorescence of neurons and astrocytes were maximally correlated with no time shift, a relationship enhanced during movement (Fig. 3d, black, r = 0.46 ± 0.06) compared with stationary periods (Fig. 3d, gray, r = 0.34 ± 0.04). We next investigated how astrocyte Ca²⁺ relates to arousal-driven neuronal activity specifically. To do this, we used principal component (PC) analysis to identify the PC of neuronal activity most correlated with arousal (arousal PC)²⁹. We found that the arousal PC of neuronal activity (Fig. 3e, top, red line) often better matched the pupil diameter (Fig. 3e, top, light gray) than the mean neuronal activity (Fig. 3e, top, dark gray), probably due to the variability in neuronal responses to arousal (Fig. 3e, bottom).

In agreement with previous work²⁹, we found that the arousal PC was usually PC1 (Fig. 3f, gray, 28/35 recordings). Furthermore, we found that average astrocyte Ca²⁺ activity (Fig. 3g, left, magenta) was better correlated with pupil diameter than average neuronal Ca²⁺ activity (Fig. 3f, left, gray). This relationship was reversed when using the arousal PC of astrocyte (Fig. 3g, right, magenta) and neuronal (Fig. 3g, right, gray) activity instead. The arousal PC consistently reflected arousal-associated activity better than using mean fluorescence for neurons (Fig. 3h, gray dots) but not for astrocytes (Fig. 3h, magenta dots), suggesting that astrocytic responses to arousal are best characterized by a consistent increase reflected in the mean fluorescence while neuronal responses are best captured by describing the variability in neuronal activity using PC analysis.

We next looked at changes in astrocyte Ca²⁺ and arousal-driven neuronal activity with arousal. We found that, with both movement (Fig. 3i, top) and stationary increases in pupil diameter (Fig. 3i, bottom), arousal-associated neuronal activity (Fig. 3i, gray) increased alongside astrocyte Ca²⁺ (Fig. 3i, magenta) and astrocyte Ca²⁺ appeared to peak when neuronal activity began to diminish (Fig. 3i). To investigate this further, we looked at the average arousal PC around the onset of all astrocyte Ca²⁺ events. We similarly found that astrocyte Ca²⁺ events occurred at the peak of arousal-associated neuronal activity (Fig. 3j, top left), and using the first derivative of the neuronal PC (Fig. 3j, bottom left), we saw astrocyte Ca²⁺ occurred at a transition point between increasing and decreasing arousal-associated neuronal activity. This was also true for astrocyte Ca²⁺ events during stationary periods (Fig. 3j, right). These results indicate that astrocyte activity is positioned to counteract the effects of arousal on the local neuronal population.

These results raised the possibility that astrocyte Ca²⁺ may respond to changes in local arousal-associated neuronal activity, rather than arousal or NE per se. To dissect the effects of neuronal activity and arousal on astrocyte Ca²⁺ we used Random Forest Regression to predict astrocyte Ca²⁺ (Extended Data Fig. 4). Using this strategy, we could explain ~85% of the variance in astrocyte Ca²⁺ fluorescence (Extended Data Fig. 4a, left). Permutation-based feature importance analysis indicated that pupil diameter was the most important predictor of astrocyte Ca²⁺, and a substantially better predictor than the next most important factor, local neuronal Ca²⁺ fluorescence (Extended Data Fig. 4a, right). These results remained consistent when using the neuronal arousal PC data (Extended Data Fig. 4b). We interpret these results as evidence that astrocytes, while responsive to local neuronal activity, are sensitive to arousal beyond the influence of nearby neurons.

Arousal-driven astrocyte Ca²⁺ is not local neuron-dependent

To further investigate how arousal and local neuronal activity drive arousal-associated astrocyte Ca²⁺, we expressed hM4Di, an inhibitory Gi-coupled Designer Receptor Exclusively Activated by Designer Drugs (DREADD)³⁰ in neurons, as well as GCaMP6f in astrocytes to image Ca²⁺ (Fig. 4a). We found no difference in the average frequency of astrocyte Ca²⁺ events from baseline following administration of either the hM4Di agonist CNO or saline, although the overall variability increased (Fig. 4b, resampled distributions from each condition). We also found no difference in astrocyte Ca²⁺ responses to movement with either saline or CNO administration (Fig. 4c,d and Supplementary Table 4). There was no difference in astrocyte Ca²⁺ responses to arousal during stationary periods, except at a high CNO concentration (5 mg kg⁻¹), where we recorded a potential enhancement of arousal-associated Ca²⁺ (Fig. 4e,f and Supplementary Table 5). These results indicate that local neuronal activity is not necessary for astrocyte Ca²⁺ responses to arousal, and that astrocytes may monitor the state of local circuit activity when responding to arousal.

**Fig. 4: Arousal-driven astrocyte Ca²⁺ is not dependent on local neuronal activity.**

NE ties astrocyte Ca²⁺ to changes in cortical state

In Fig. 3, we showed that astrocyte Ca²⁺ events occur before reductions in arousal-associated neuronal activity. We wondered whether astrocyte Ca²⁺ was similarly related to activity changes in the broader cortical circuit. To answer this question, in a subset of mice expressing Ca²⁺ indicators in neurons (hysn-GCaMP6f) and astrocytes (GFAP-jRGECO1b), we implanted low-impedance electrodes bilaterally into the visual cortex (Fig. 5a). Using this methodology, we recorded the LFP (Fig. 5b, top), which unlike our spatially restricted imaging window, reflects the integrated electrical inputs and activity of the visual cortex and surrounding brain regions³¹. In these LFP recordings, we observed spontaneous transitions between cortical states dominated by LF power (2–7 Hz; running-evoked 7–10 Hz was excluded) and those dominated by HF (70–100 Hz) power, indicative of synchronous and desynchronized cortical states respectively (Fig. 5b)¹. We noticed that, during transitions to a low-arousal state following movement, astrocyte Ca²⁺ peaked as LF power increased (Fig. 5c, left) and HF power decreased (Fig. 5c, right).

**Fig. 5: NE-dependent astrocyte Ca²⁺ occurs at the crux of cortical state changes.**

To investigate this relationship, we cross-correlated astrocyte and neuronal Ca²⁺ with LFP power, and with the derivative of LFP power. We found that astrocyte cross-correlations with LFP power were right-shifted from the equivalent neuronal cross-correlations (Fig. 5d,e). When looking at LF power specifically (Fig. 5d, left, magenta), we found that astrocyte Ca²⁺ preceded increases in LF power (Fig. 5d, left, magenta). In contrast, neuronal Ca²⁺ was only significantly positively correlated with LF power with zero lag (Fig. 5d, left, gray), indicating that LF power and neuronal population activity are coordinated. These findings were confirmed when examining the derivative of LF power. We found that changes in LF power were maximally correlated with astrocyte activity without any lag (Fig. 5d, right, magenta), while LF power changes preceded neuronal activity (Fig. 5d, right, gray). Our results demonstrate that, while neuronal Ca²⁺ is correlated with LF power, astrocyte Ca²⁺ is correlated with changes in LF power.

We found a similar relationship when using HF power, although inverted about the y axis. HF power preceded both neuronal and astrocyte Ca²⁺, with neurons maximally negatively correlated with zero time lag (Fig. 5e, left). In addition, both astrocyte and neuronal Ca²⁺ were negatively correlated with the derivative of HF power, but for astrocytes this relationship was true both when HF power preceded or had no lag with astrocyte Ca²⁺ (Fig. 5e, right). These results indicate that, while HF power precedes both neuronal and astrocyte Ca²⁺, neuronal Ca²⁺ is inversely related to absolute HF power while astrocyte Ca²⁺ is inversely related to changes in HF power. Overall, this analysis demonstrates that neuronal Ca²⁺ activity reflects absolute LFP power while astrocyte Ca²⁺ is instead tied to changes in LFP power.

We next asked whether arousal-driven astrocyte Ca²⁺ was generally tied to changes in LFP power by computing an astrocyte Ca²⁺ event-triggered spectrogram. We found that astrocyte Ca²⁺ signaling occurred at the crux of a transition from a HF-dominated cortical state to one dominated by LF power (Fig. 5f). This relationship was also seen in the LFP power of the contralateral cortex (Fig. 5g), indicating that astrocyte Ca²⁺ is tied to the transition of neuronal activity from a high- to low-arousal state not just at a local level (Fig. 3j), and at the level of the nearby cortex (Fig. 5f), but also to bi-hemispheric changes in cortical state.

On the basis of our identified changes in arousal and NE as major drivers of spontaneous changes in astrocyte Ca²⁺ (Figs. 1 and 2), we hypothesized that NE was necessary for the link between astrocyte Ca²⁺ and cortical state transitions. In particular, we hypothesized that Adra1 receptors might underlie this relationship due to their importance for astrocyte physiology^14,16,19,32. We pharmacologically blocked Adra1 receptors using Prazosin (5 mg kg⁻¹, intraperitoneal (i.p.)) and found that the relationship between astrocytes and cortical state was reduced (Fig. 5h,i), confirming that NE signaling links astrocyte Ca²⁺ to cortical state changes.

Enrichment of Adra1a mRNA in visual cortex astrocytes

We next sought to determine which Adra1 receptors were most likely to underlie the connection between astrocytes and cortical state. We analyzed a previously published dataset that profiled the ribosome-associated messenger RNA expression of astrocytes in the mouse visual cortex³³. We looked at adrenergic receptor expression in the adult (P120) dataset and confirmed astrocytic expression of most adrenergic receptors, although astrocytes showed little Adra2b, Adrab2 and Adrab3 in this dataset (Extended Data Fig. 5a, left). When analyzing the astrocytic expression of adrenergic receptors relative to an input control, the relative expression of Adra1a mRNA was highest and greater than the input control, indicating that Adra1a mRNA may be preferentially enriched in visual cortex astrocytes (Extended Data Fig. 5a, right).

We also used spatial transcriptomics to assess the mRNA expression of Adra1a, Adra1b and Adra2a (Extended Data Fig. 5b). Using LaST map single-molecule in situ hybridization (smFISH)³⁴, we saw striking heterogeneity throughout the brain (Extended Data Fig. 5b, left). All three receptors showed cortical expression in astrocytes, as delineated by expression of the astrocyte-specific mRNA SlcA3 (GLAST), and in non-astrocytic cell types (Extended Data Fig. 5b, right). We quantified mRNA spots per astrocyte in the visual cortex for each of these receptors and found higher expression at deep layers of cortex (Extended Data Fig. 5c). This was particularly true for Adra2a and Adra1b receptors, while Adra1a receptors also showed higher expression in intermediate layers of cortex (Extended Data Fig. 5c), where the effects of cortical synchrony are particularly important for perception³⁵.

Astrocyte Adra1a receptor signaling shapes neuronal activity

On the basis of pharmacological and transcriptomic data, as well as its activation of astrocytes in other contexts^15,16,17,36, we identified Adra1a receptors as a likely connector between astrocytes and cortical state. To test how Adra1a receptors modulated cortical activity, we injected the Adra1a receptor agonist A61603 (10 μg kg−1, i.p.)—which stimulates astrocytes in acute cortical slices³²—while recording cortical astrocyte and neuron Ca²⁺ activity (Fig. 6a). We saw robust responses within minutes of A61603 administration (Fig. 6b). Astrocyte Ca²⁺ was more homogeneous (Fig. 6c) and elevated (Fig. 6d, left), while neuronal Ca²⁺ showed a general reduction in baseline Ca²⁺ (Fig. 6d, right), while exhibiting similar oscillatory activity as astrocytes. To better quantify these neuronal changes, we took the power spectrum of each neuron and separated the activity into slower (<0.05 Hz, Fig. 6e, left) and faster fluctuations (>0.05 Hz, Fig. 6f, left). We found that after A61603 administration there was more power in <0.05 Hz fluctuations (Fig. 6e, right) and less power in >0.05 Hz fluctuations (Fig. 6f, right). Our results indicate that A61603 administration, while increasing the coordination and level of astrocyte Ca²⁺ activity, broadly inhibits and alters the pattern of neuronal activity.

**Fig. 6: Adra1a receptors modulate basal neuronal activity and neuronal population responses to arousal.**

To determine how astrocytic Adra1a receptors specifically affect cortical neuronal activity, we generated a mouse line with LoxP sites flanking Adra1a, allowing for Cre-specific deletion of the receptor (Extended Data Fig. 6). We injected an astrocyte-specific Cre-GFP virus into the cortex of Adra1a^fl/fl mice (or wild-type littermate controls), and performed neuronal Ca²⁺ imaging while recording movement and pupillometry (Fig. 6g). In agreement with previous work³⁷, we saw astrocyte-specific Cre expression which colocalized with the astrocytic marker S100β, but not the neuronal marker NeuN (Extended Data Fig. 7). Overall activity in these mice showed more neuronal Ca²⁺ events (Fig. 6h), as well as less power in <0.05 Hz Ca²⁺ fluctuations (Fig. 6i) and more power in >0.05 Hz Ca²⁺ fluctuations (Fig. 6j), the inverse of stimulating Adra1a receptors using A61603 (Fig. 6d–f). These results suggest NE signaling through astrocyte Adra1a receptors is a major pathway for modulating the level and pattern of neuronal activity. To ensure these results were not an artifact of imaging rate, we recorded neuronal activity using resonant galvanometers and saw the same results (Extended Data Fig. 8).

We next determined whether astrocyte Adra1a receptors impacted arousal-related neuronal activity. As before (Fig. 3i), we assessed how arousal-related neuronal activity changed either with movement (Fig. 6k) or stationary pupil dilation (Fig. 6l). In this dataset, the distribution of durations during movement bouts (Extended Data Fig. 9a) and stationary pupil dilations (Extended Data Fig. 9b) was wide, and differed between Adra1a^fl/fl mice and wild-type controls. To account for this, we analyzed only arousal-related neuronal activity during the movement (Extended Data Fig. 9b) or pupil dilation event (Extended Data Fig. 9c), including a ~1 s offset to account for a delayed response in neuronal activity. In both cases, Adra1a^fl/fl mice displayed enhanced arousal-related neuronal activity (Fig. 6k,l). This increase could not be explained by an overall stronger neuronal connection to arousal; Adra1a^fl/fl neuronal activity was less correlated with pupil diameter than that in wild-type littermates (Fig. 6m). These results suggest that NE signaling through astrocyte Adra1a receptors regulates both the magnitude and shape of neuronal responses to arousal.

NE signaling via astrocyte Adra1a modulates cortical state

On the basis of the Adra1a-dependent relationship between astrocytes and cortical state, we hypothesized that, in contrast to the desynchronizing effect of NE on the cortex generally¹¹, signaling through astrocytic Adra1a receptors would lead to cortical synchrony. We again injected the Adra1a receptor agonist A61603 (1 μg kg⁻¹, i.p.) while recording cortical LFP (Fig. 7a). Compared with saline control, A61603 treatment resulted in high-amplitude LFP fluctuations (Fig. 7b) reflected by increased LF power (Fig. 7c,d). In contrast, we saw no significant difference in HF power (70–100 Hz, Fig. 7e). These effects show that, contrary to the role of NE more broadly^10,11,25, stimulation of Adra1a receptors increases cortical synchrony.

**Fig. 7: Genetic removal of astrocyte Arda1a impairs cortical resynchronization after arousal.**

However, the manipulation using A61603 was body-wide due to i.p. injection and affected all Adra1a-expressing cells. On the basis of our Ca²⁺ imaging results (Fig. 6g–m) in Adra1a^fl/fl mice, we wondered whether astrocytic Adra1a receptor signaling might modulate cortical state. To test this, we used the same receptor knockout strategy and recorded cortical LFP (Fig. 7f). In Adra1a^fl/fl mice, the cortical LFP was generally higher amplitude than in controls (Fig. 7g), with higher power across all frequency bands (Fig. 7h, left), and higher total LFP power (Fig. 7h, right). These results further support the idea that NE signaling to astrocytes modulates overall cortical neuronal activity.

We also hypothesized that removing astrocytic Adra1a receptors would effect NE-related cortical state transitions. We focused on times when mice stopped moving (Fig. 7i, top row), and found Adra1a^fl/fl mice had increased HF power during movement (Fig. 7i, middle row; Fig. 7j) and less LF power in the Adra1a^fl/fl mice compared with wild-type mice (Fig. 7i, bottom row; Fig. 7k). Taken together, these data show that selectively removing a NE receptor from astrocytes alters arousal-related cortical state changes, enhancing the desynchronizing effects of arousal and reducing resynchronization afterwards.

Discussion

Understanding awake cortical state regulation is crucial for understanding perception, attention and behavior¹. Most research examining cortical states has focused on arousal mechanisms—such as release of NE—that desynchronize the cortex and increase sensitivity to external stimuli^6,7,24. However, we found that NE (Fig. 8, green) not only desynchronizes cortical neuronal activity, but also leads to cortical synchrony by activating astrocytic Adra1a receptors. This model identifies astrocytes as key regulators in the awake cortex, acting as a feedback mechanism for arousal-associated desynchrony (Fig. 8).

**Fig. 8: Model of astrocyte regulation of arousal-associated cortical state.**

Astrocytes regulate arousal-associated cortical circuits

We found that astrocytes responded more sensitively to arousal (Fig. 1) and NE (Fig. 2) than previously recognized^14,15,16,17. Arousal-associated astrocyte Ca²⁺ activity occurred at the crux of reductions in arousal-associated neuronal activity (Fig. 3) and bi-hemispheric cortical state changes (Fig. 5), and was not dependent on local neuronal activity (Fig. 4). These results suggest direct NE signaling to astrocytes acts as a separate neuromodulatory pathway. To support this, we found that NE signaling through astrocyte Adra1a receptors regulates overall neuronal activity (Fig. 6h–j), the response of neurons specifically to arousal (Fig. 6k–m), and arousal-associated cortical states (Fig. 7f–k). Notably, we find that astrocytes detect relative changes in arousal, suggesting they act as a feedback mechanism without working against the arousal system more generally. This role may be particularly useful in the context of heterogeneous cortical NE dynamics. We observed extracellular fluctuations in NE²⁶ on a timescale of seconds, similar to the reported activity patterns of the underlying locus coeruleus axons²⁸, as well as much slower signaling on the order of tens of seconds to minutes (Fig. 2c). Thus, change detection may allow astrocytes to ignore the tonic fluctuations of cortical NE and sensitively respond to phasic increases in NE.

Astrocyte Ca²⁺ is poised to regulate neuronal activity

Of course, astrocytes are not the sole detectors of cortical NE. Neurons, among other cell types^12,13, also show strong responses to arousal^2,3,7,24 and NE¹⁰. Understanding the relative effects of arousal on neuronal and astrocytic activity is crucial to understanding how these cellular populations contribute to arousal-mediated cortical state changes. Our results are notable in the context of previous research placing astrocyte activity downstream of neuronal activity, or to driving changes in neuronal activity on long timescales^32,38,39. In contrast, our data show that arousal-associated astrocyte Ca²⁺ signaling occurs at the crux of the neuronal response to arousal, co-occurring with transitions in arousal-associated neuronal activity (Fig. 3). This relationship was also seen more broadly, with astrocyte Ca²⁺ occurring when bi-hemispheric arousal-associated desynchronized electrical activity decreased and cortical synchrony increased (Fig. 5). Furthermore, we found that NE-specific astrocyte Ca²⁺ signaling, and its relationship to arousal, was not dependent on the activity of nearby neurons, suggesting that astrocytes act directly downstream of neuromodulatory signaling to modulate neuronal activity (Fig. 4). These findings support a model in which astrocytes dynamically respond to NE to modulate neuronal activity on the physiological timescales of arousal.

Receptor-specific relationship between NE and cortical state

Our work positions cortical astrocytes as drivers of neuronal synchrony on multiple scales in the wake state, and suggests new roles for NE in cortical state regulation. In contrast to the generally desynchronizing role of NE¹¹, we found that specific pharmacological stimulation of the Adra1a receptor reduced cortical neuronal activity (Fig. 6d, right) and altered its pattern (Fig. 6e, f). Adra1a receptor stimulation also led to cortical synchrony increases without altering HF power (Fig. 7a–d). This relationship was initially hypothesized by early pharmacological work⁴⁰, but to our knowledge has not been validated nor widely accepted. Our findings suggest that even within the same family, activation of different NE-receptor subtypes may have profoundly different effects on cortical state and by extension arousal, attention and behavior, which may help explain disparate effects of adrenergic signaling on perception and cognition^41,42.

What is arousal?

Here we used pupil diameter as an external read-out for a complicated set of biological processes that influence neural activity brain-wide^23,25,43,44. While we focused on NE, many neuromodulators including acetylcholine^28,45 and serotonin⁴⁶ vary with arousal and modulate cortical state. Fully describing how neuromodulatory inputs affect astrocyte and neuronal activity will be vital for a complete description of cortical state regulation. Other internal states, such as motivation, interact with arousal and may involve neuromodulatory systems, but are not fully captured with the present arousal metrics. Here we were interested in pupil diameter and movement as external readouts of NE-related states, but future work should investigate the contributions of other internal states, using tools such as facial motion energy²⁹, to further dissect astrocyte function in the context of internal states.

While our work establishes a role for astrocytes in mediating arousal-associated changes in neuronal activity and cortical state, we did not directly link these changes to behavioral outcomes. Future studies should focus on how arousal-associated astrocyte Ca²⁺ influences the general perceptual effects of arousal⁷ and extend these findings to the specific effects of attention^1,35,43,47. It will also be important to examine how sensory stimuli-driven^48,49,50 and arousal-driven astrocyte Ca²⁺ signaling leads to changes in behaviorally relevant neuronal population activity and cortical synchrony.

How does arousal modulate different cellular subpopulations?

To expand the impact of our findings, the experiments described here could be extended to dissect the role of specific subpopulations of neurons and astrocytes, or types of Ca²⁺ activity in these populations. In this work, we analyzed Ca²⁺ imaging data as homogeneous cellular populations, in part due to an acquisition rate (~2 Hz) necessary for imaging astrocyte populations on a wider scale (0.5 mm²). However, there may be differences between the population-level Ca²⁺ activity we identified and fast Ca²⁺ responses to arousal.

Both astrocytes^34,51 and neurons⁵² can be split into multiple subpopulations, and neuronal subtypes—in particular inhibitory interneuron subtypes—have unique relationships to arousal and cortical state⁵³. In our work, we used PC analysis to account for heterogeneous neuronal responses to arousal, and we found Adra1a-mediated astrocytic modulation of neuronal responses to arousal. However, we did not identify any specific effects on neuronal subtypes. On the basis of the overall reduction of neuronal activity with A61603 injection (Fig. 6d, right), and the overall increase in neuronal activity (Fig. 6h) and LFP power with knockout of astrocytic-Adra1a receptors (Fig. 7h), our work suggests astrocytic modulation of cortical inhibition is a likely candidate for the downstream circuit mechanisms controlling cortical state, as has been proposed in other contexts^36,54. Cortical interneurons have been shown both to play a critical role in the control of cortical state^45,55,56, and to be particularly sensitive to their electrostatic extracellular environment⁵⁷. Neuromodulator-driven astrocytic modulation of the extracellular environment⁵⁸ and inhibitory regulation of nearby neurons¹⁸ may act as prime loci of control for cortical circuit activity and state.

Within our smFISH dataset (Extended Data Fig. 5) and in published single-cell data⁵¹, cortical astrocytes show variable expression of neuromodulatory receptors, and some astrocytes lack any transcripts for Adra1a. This suggests molecularly distinct, and potentially functionally distinct, subpopulations of cortical astrocytes with differential sensitivity to arousal-related neuromodulators. It also suggests that gap junctions, ATP signaling and other mechanisms^59,60 that create networks within the astrocyte syncytium may be important for arousal-associated astrocyte activity and modulation of cortical circuits. Identifying the specific relationships among subpopulations of astrocytes and neurons, and determining the manner and magnitude of their modulation by arousal, will be important for understanding what constitutes a cortical circuit and how cortical circuit activity is regulated. Furthermore, much like the effects of arousal, these relationships may be layer³⁵ or cortical area⁴⁴ specific. Future study of cellular subpopulation-specific effects on cortical state, particularly in additional cortical regions and in combination with behavioral assessments, will provide a richer appreciation of the cellular mechanisms that regulate arousal-associated cortical state.

Methods

Animals

All procedures were carried out in accordance with protocols approved by the University of California, San Francisco Institutional Animal Care and Use Committee. Animals were housed and maintained in a temperature-controlled environment on a 12 h light–dark cycle, with ad libitum food and water. Male and female mice were used whenever available. All imaging/electrophysiology experiments were performed at the same time each day. Adult C57BL/6 mice, Adra1a^fl/fl mice or Adra1a wild-type mice aged 1–6 months at time of surgery were used. For experiments involving Adra1a ^fl/fl mice, the experimenter was blind to animal genotype before surgery, recording and analysis.

Generation of Adra1a^fl/fl mice

The Mouse Biology Program at the University of California, Davis, constructed the mouse. A floxed FLAG-α1A knock-in vector was made using standard methods as follows: The targeting vector contained a 5′ arm of 5.4 kb and a 3′ arm of 5.5 kb. Lox P sites were placed upstream and downstream of the α1A-AR gene first coding exon. A Kozak sequence and single FLAG tag were upstream, and the neomycin-resistance gene was downstream.

The vector was electroporated into the C57BL/6N ES cell line JM8.F6. The resulting ES cell clones were screened by long-range PCR for loss of the native allele and homologous recombination, and containing a single copy of the plasmid by LoxP PCR. Two clones passed these screens and had normal chromosome counts. Both ES clones were microinjected into blastocysts, and transferred to embryonic day 2.5 stage pseudo-pregnant recipients. The resulting chimeras were screened for percent ES cell derived coat color, and those greater than 50% were mated to C57BL/6N females. Germline heterozygous mice were produced, and mated to MMRRC strain C57BL/6-Tg(CAG-Flpo)1Afst/Mmucd, RRID: MMRRC:036512-UCD, for excision of the Neo selection cassette. Neo-excised mice were identified via PCR, and mated further to C57BL/6J mice to remove the FLP transgene. Mice were continued in C57BL/6J.

Routine PCR genotyping used 5′-gcttcctcaggctcacgtttcc and 3′-gccttagaaatgttcacctgtgc primers upstream and downstream of the LoxP site (Extended Data Fig. 6). These mice are available upon request from the corresponding author.

Surgical procedures and viral infection

Mice were administered dexamethasone (5 mg kg⁻¹, subcutaneous) at least 1 h before surgery, and anesthetized using 1.5% isoflurane (Patterson Veterinary Supply, 78908115). After hair removal and three alternating swabs of 70% ethanol (Thermo Fisher Scientific, 04-355-720) and Betadine (Thermo Fisher Scientific, NC9850318), a custom-made titanium headplate was attached to the skull using cyanoacrylate glue and C&B Metabond (Parkell, S380). If recording LFP, 0.5 mm burr holes were made bilaterally over the visual cortex, and bilaterally over the cerebellum for reference, and a ~200-µm-diameter perfluoroalkoxy-coated tungsten wire (A-M Systems, 796500) was implanted -0.2 mm into each hole and secured with Metabond. For imaging and LFP experiments, a 3 mm craniotomy was made over the right visual cortex and the right visual cortex burr hole was made lateral to the craniotomy.

For viral infection, 400–800 nl total volume of the following viruses were injected alone, or in combination by premixing the solutions before aspiration: AAV5.GfaABC₁D.GCaMP6f.SV40 (Addgene, 52925-AAV5), AAV9.hSyn.NE2h (WZ Biosciences, YL003011-AV9), AAV9.GfaABC₁D.jRGECO1b (Vigene, custom-ordered), AAV9.Syn.GCaMP6f.WPRE.SV40 (Penn Vector Core, AV-8-PV2822), AAV2.hSYN.hM4D(Gi).mCherry (Addgene, 50475-AAV2) and AAV5.GFAP(0.7).EGFP.T2A.iCre (Vector Biolabs, VB1131).

Injections were made through a glass pipette and UMP3 microsyringe pump (World Precision Instruments) into one or two locations in the right visual cortex at coordinates centered on +2.5 mm medial/lateral, +0.5 mm anterior/posterior and −0.3 dorsal/ventral from lambda. After allowing at least 10 min for viral diffusion, the pipette was slowly withdrawn and a glass cranial window implanted using a standard protocol⁶¹.

For in vivo fiber photometry recordings, GRAB_NE2h (AAV9.hSyn.NE2h) and astrocytic jRGECO1b (AAV9.GfaABC₁D.jRGECO1b) were expressed via viral vectors in C57Bl/6 mice. A 1-mm-diameter craniotomy was made over the PFC (+1.7–1.8 mm rostral, +0.5 mm lateral from bregma), and viral vectors were delivered to −2.3 to 2.4 mm ventral. A fiber optic cannula (Mono Fiberoptic Cannula, 400 µm core, 0.66 NA, 2.8 mm length, Doric Lenses) was then lowered to −2.3 mm ventral and secured in place using dental cement.

Immunohistochemistry and in situ hybridization

Immunohistochemistry

Following in vivo experiments, mice were overdosed on isoflurane and then perfused transcardially with phosphate-buffered saline (PBS, Sigma-Aldrich P3813) followed by 4% paraformaldehyde in PBS (Santa Cruz Biotechnology, CAS 30525-89-4). Brains were removed and postfixed overnight in 4% paraformaldehyde, followed by cryoprotection in 30% sucrose in PBS for 2 days at 4 °C. Brains were snap frozen in dry ice and stored at −80 °C until adhesion onto a sectioning block with Optimal Cutting Temperature Compound (Thermo Fisher Scientific, 23-730-571). Forty-micrometer coronal sections were taken on a cryostat and stored in cryoprotectant at −20 °C until immunohistochemistry was performed. For immunohistochemistry, free-floating sections spanning the rostral–caudal axis were selected and washed with 1× PBS for 5 min three times on an orbital shaker, followed by permeabilization with 0.01% PBS-Triton X for 30 min. Sections were then blocked with 10% NGS (Sigma-Aldrich, S26-100ML) for 1 h. Immediately after, sections were incubated with chicken α-GFP (1:3,000, Abcam, ab13970) and either rabbit α-NeuN (1:1,000, EMD Millipore, ABN78) or mouse α-NeuN (1:1,000, Millipore Sigma, MAB377) and rabbit α-S100B (1:500, Millipore Sigma, SAB5500172). Sections were then washed with 1x PBS for five minutes three times on a shaker, followed by secondary incubation with Thermo Fisher Scientific goat α-chicken Alexa Fluor 488 and either goat α-rabbit Alexa Fluor 405 or goat α-mouse Alexa Fluor 405; goat α-rabbit Alexa Fluor 555 for 2 h at 20 °C on a shaker. Sections were washed again with 1× PBS for 5 min three times, then mounted and cover slipped with Fluoromount-G. For whole section examples, images were taken using a Keyence BZ-X800 fluorescence microscope to assess viral spread. Then, 2× images were acquired and the images were computationally stitched with Keyence Analysis Software. For cell counting 60× z-stacks were captured on a spinning-disk confocal (Zeiss). Slides were oil-immersed. The Fiji plugin Cell Counter was used to quantify the number of GFP⁺, NeuN⁺ and GFP⁺/NeuN⁺ cells to determine colocalization. Each animal had two sections, with each section having six distinct field of views containing 20 z-planes. A cell was considered GFP⁺ when signal was confined to soma and processes, and GFP⁺/NeuN⁺ when cells exhibited merged signals.

Single-molecule fluorescent in situ hybridization

Single-molecule, fluorescent in situ hybridization data collection was performed using LaST map smFISH as previously described³⁴. Astrocyte cell boundaries were segmented using an ilastik pixel classifier and a customized watershed segmentation. To preserve processes of astrocyte and not cut them off from the DAPI signals, a pixel classifier was trained by using only large Gaussian filters (5/10 pixels) during the feature extraction step. As a result, astrocyte processes were detected with fewer splits. Pixel classification carves out only non-background pixels from the image and does not identify the astrocyte boundaries separating neighboring cells, that is, instance segmentation. To address this problem, we first used CellPose to identify all nuclei from the image. Subsequently, the centroids of all nuclei were extracted and used as the seeds to generate astrocyte boundaries between adjacent cells using a watershed segmentation. As a result, astrocytes with touching processes were separated. Finally, as non-cell debris might remain in the segmentation image, we further used an object classification workflow in ilastik to remove them.

All mRNA signals were detected using a Python package called TrackPy using five pixels as diameter and 96 as percentile threshold. These detected spots were then assigned to each astrocyte that they sat within by using a Python package called shapely. Due to the tissue damage that occurred during the sample preparation step, only seven cortex surface areas from all sections were eligible for downstream processing to ensure the comparability between tissue sections. For all these regions, both white matter boundaries and the superficial cortex area were manually annotated. For each cell, the distance to both the white matter boundary (d_WM) and distance to the cortex surface (d_ep) were calculated. The relative cortical depth (D_relaCortex) was thus defined as:

$$D_{{\mathrm{relaCortex}}} = \frac{{d_{{\mathrm{ep}}}}}{{d_{{\mathrm{ep}}} + d_{{\mathrm{WM}}}}}$$

Recording setup

Animals were given at least 1 week after surgery for recovery and viral expression. They were then habituated on a custom-made circular running wheel over at least 2 days, and for a cumulative time of at least 2.5 h, before experimental recordings began. After habituation, mice were head-fixed on the wheel and movements were recorded by monitoring deflections of colored tabs on the edge of the wheel using an optoswitch (Newark, HOA1877-003).

Pupillometry

Pupil recordings were made using a Genie near-infrared camera (1stVision, M640) and a telescopic lens (Thorlabs, MVL50TM23), and acquired at 30 Hz using the MATLAB Image Acquisition toolbox. A small monitor (Amazon, B06XKLNMW3) showing a consistent teal background color (RGB: 0,1,1) was placed by the mouse to allow for observation of the full range of pupil dynamics in an otherwise dark room. For experiments without 2P illumination, a near-infrared light (Amazon, B00NFNJ7FS) was used to visualize the pupil.

2P imaging

2P imaging was performed on a microscope (Bruker) with two tunable Ti:sapphire lasers (MaiTai, SpectraPhysics) and a Nikon 16×, 0.8 numerical aperture water-dipping objective with a 2× optical zoom (frame rate 1.7 Hz, field of view 412 µm², resolution 512 × 512 pixels). A 950 nm excitation light with a 515/530 emission filter was used to image green-emitting fluorophores, and 1,040 nm light with a 605/615 emission filter was used to image red-emitting fluorophores. Recordings lasted from 10 min to 1 h.

Electrophysiology

Visual cortex LFP was recorded at 1 kHz and subtracted from the ipsilateral cerebellar LFP before 1 kHz amplification (Warner Instruments, DP-304A) and acquired using PrairieView (Bruker) or PackIO⁶².

In vivo pharmacology

Recordings were taken before and after saline, Prazosin–HCl (5 mg kg⁻¹ Sigma-Aldrich, P7791-50MG), A61603(1 µg kg⁻¹ or 10 µg kg⁻¹, Tocris, 1052), or clozapine N-oxide (CNO, 1 mg kg⁻¹ or 5 mg kg⁻¹, Tocris, 4936) were injected intraperitonially while animals remained head-fixed on the wheel, to ensure post-treatment recordings were comparable with baseline measures.

Fiber photometry

Dual-color fiber photometry recordings were performed on a Tucker-Davis Technologies RZ10X processor with 405, 465 and 560 nm LEDs. LED drivers were modulated such that light power was approximately 15 μW for 405 nm and 20 µW for 465 nm and 560 nm wavelengths at the tip of the light path. Animals were recorded in a freely moving arena in which the mouse was able to move in all directions, after coupling to low-autofluorescence fiberoptic patchcords connected to photosensors through a rotary joint (Doric Lenses). Fluorescence signals were recorded for 10 min, during which tail lifts were performed every 2 min. For a tail lift stimulation, the experimenter held and lifted the tail of the animal until its hind paws disconnected from the ground; after that the tail was released. With this experimental paradigm, no pain or harm is caused to the animal.

Statistics and reproducibility

All data analysis was done in MATLAB unless otherwise indicated. No statistical methods were used to predetermine sample sizes, but they are similar to previous reports^14,28. Box plots are shown with the central mark indicating the median and the bottom and top edges of the box indicating the 25th and 75th percentiles, respectively. Whiskers extend to the most extreme data point or within 1.5 times the interquartile range (IQR) from the bottom or top of the box, and all other data are plotted as individual points, as listed in the figure legends. For statistical comparisons, nonparametric tests were used, or where indicated, normality was assumed but not formally tested, and t-tests were used. HB was performed on the basis of a MATLAB implementation (https://github.com/jenwallace/Hierarchical_bootstrap_Matlab) of the methodology, and used to reduce the statistical error rate of comparisons while retaining statistical power⁶³. All multiway comparisons were adjusted for using Tukey–Kramer correction. No data were excluded from analyses except for the following (not predetermined): In hSYN-hM4Di experiments, outliers were excluded across all conditions from small stationary responses to avoid confounding effects from other influences on astrocyte Ca²⁺, as described in methods. For in vivo pharmacology experiments, electrical artifacts in band power were excluded before analysis, as described in methods. Samples were allocated into experimental groups by cell-type expression of each individual fluorescent sensor. Only adult animals (1–6 months of age) were used in experiments, and both males and females were used and randomly selected. For imaging and electrical recordings of spontaneous activity, blinding was not relevant because cell-type viral expression is evident from expression pattern. For in vivo pharmacology, blinding was not possible because control recordings were taken before treatment recordings to avoid confounding the treatment effects. For Adra1a^fl/fl mice, the experimenter was blinded to genotype before data collection and analysis.

Speed calculations

To compute wheel speed, a detected break in the optoswitch circuit was determined when the absolute value of the derivative of the raw voltage trace was at least 2 s.d. above the mean. For recordings with very little movement (s.d. <0.1), this threshold generated false positives so a set threshold of 0.1 was used. The number of breaks in the optoswitch circuit per second was then calculated, and using the circumference and number of evenly spaced colored tabs at the edge of the wheel, the wheel speed was determined and used for all subsequent analyses using speed. Movement periods were defined by wheel speed ≥10 cm s⁻¹, and movement bouts that were separated by ≤2 s were considered one event. To ensure that movement-related dynamics were not included in stationary analysis, data were excluded from at least 10 s around identified movement periods.

Pupillometry

Following acquisition, pupil data were processed through a Python function that used contrast detection to identify the edges of the backlit pupil from the sclera and fit an ellipse whose major radius was taken as the pupil diameter. The diameter was then low-pass filtered to 0.5 Hz and normalized to a range between 0 and 1 to give pupil diameter as percent maximum. The pupil derivate was normalized to the acquisition rate (30 Hz) to compute pupil phase and to determine the phase of astrocyte Ca²⁺ events and the cross-correlations with GRAB_NE. Stationary arousal dilations and constrictions were identified by the sign of the calculated pupil derivative, and only changes in pupil diameter >10% were used for subsequent analyses.

LFP

All spectral analysis was done using Chronux⁶⁴. Raw LFP data were visually inspected to confirm useable signal was present, and then 60 Hz noise was filtered out and drifting baselines were compensated for using linear fitting. LFP power for frequency bands was computed using built-in Chronux functions with a time bandwidth of 2.5 and two tapers, no frequency padding and 5 s moving windows. For changes in LFP band power around arousal or astrocyte Ca²⁺ events, the median band-limited power was obtained and then normalized to the median band-limited power in the event-triggered time window to get relative band power. The median power before an event onset versus after was combined for each recording, and for Adra1a^fl/fl mice, the ratio between the two was computed and compared with Adra1a wild-type mice. For changes in LFP power after A61603 administration, the band-limited power for saline and drug data was calculated, outliers were removed to avoid contamination by recording artifacts, and then this power was normalized to the band-limited power for the respective baseline recording.

For total power in Adra1a^fl/fl and control mice, no baseline correction was done to avoid skewing the analysis. Spectrum power was calculated by concatenating recordings from all mice of each genotype and computing average or individual spectra with a time bandwidth of six and eight tapers to increase accuracy. The total power was then computed by summing across all frequency bands. Relative power was computing by dividing the spectrum from each mouse by its total power, and relative band power was computed by summing the power from each frequency band and dividing by the total power.

Ca²⁺ imaging

Astrocyte Ca²⁺ events and fluorescence was extracted using the AQuA software analysis package²². For dual-color imaging with neuronal Ca²⁺ indicators, particular care was taken to avoid AQuA detection of neuronal activity; the s.d. of the neuronal channel was taken in FIJI and a mask was created in AQuA to exclude areas of high neuronal activity and soma from analysis. Astrocyte events were included only if they had an area greater than 10 µm, lasted for at least two frames and had an AQuA P value <0.05. To obtain the average astrocyte Ca²⁺ fluorescence, the compensated fluorescence traces that account for spatially overlapping events were taken, normalized to their maximum value and then averaged together.

Neuronal Ca²⁺ events and fluorescence were extracted from neuropil background semi-automatically using Suite2P⁶⁵. We identified Ca²⁺ events by taking identified spikes in the Ca²⁺ fluorescence data and thresholding them for only the largest (>3 s.d. over the mean) events. To calculate the average neuronal Ca²⁺ fluorescence, the trace from each neuron was normalized to its maximum value and averaged together.

Machine-learning based analysis of input contribution to astrocyte Ca²⁺

As an alternative to assess the contribution of biological inputs on astrocyte Ca²⁺, data from dual-color Ca²⁺ imaging were used to train a machine-learning model to predict average Ca²⁺ activity. To include LFP data, the spectrogram data from each LFP recording were decomposed using PC analysis based on the eigenvectors from the ipsilateral recording that accounted for the largest proportion of the variance. The PC1 in this data corresponded to cortical synchrony, with positive weights for HF and negative weights for low frequencies, matching a previous report but with inverse sign⁶⁶. LFP PC1 for ipsilateral and contralateral recordings, as well as speed, pupil diameter and average Ca²⁺ fluorescence for neurons and astrocytes, was then z-scored and resampled to 10 Hz before being concatenated. This dataset was then imported into Python for machine learning analysis using the SciKit-learn toolbox. For analysis, randomly generated data were added for comparison, and rows without both ipsilateral and contralateral LFP recordings were excluded from subsequent analysis. Average astrocyte Ca²⁺ data were used as the target dataset and data were split into training (80%) and testing (20%) before classification using a random forest regression model. The model was validated using the R² between the predicted average astrocyte Ca²⁺ fluorescence from the model and the actual average astrocyte Ca²⁺ data of the test set. Permutation testing of the predictors was then used to determine their relative contributions to model prediction.

GRAB_NE analysis

In GRAB_NE imaging data, we observed background fluorescence fluctuations that we thought might arise from hemodynamic artifacts. To ensure the data reflected the NE signal, hemodynamic artifacts in the data were removed by a custom-designed, data processing pipeline. The predominant hemodynamic artifact in the data was assumed to reflect fluctuating hemoglobin levels altering brain absorptivity causing an attenuation of light. As such, the signal from each pixel could be modeled as

$$Y_k\left( t \right) = F_k\left( t \right) \cdot e^{ - \mu _k\left( t \right)X} + N\left( t \right) = F_k\left( t \right) \cdot e^{h_k\left( t \right)} + N\left( t \right)$$

where k is index of pixel, Y_k(t) and F_k(t) are respectively the observed curve and the real fluorescence of kth pixel, μ_k(t) is the absorption coefficient for kth pixel, X is the path distance, the term $e^{h_k\left( t \right)}$ represents the intensity attenuation, and N(t) is the noise. In our data, the identified hemodynamic signal across pixels was approximately synchronous but varied in magnitude; thus, the attenuation of one pixel can be represented by another, that is $e^{h_j\left( t \right)} = e^{ah_k\left( t \right) + b}\left( \,{j \ne k} \right)$. On the basis the findings above, we designed the following pipeline:

1.
We selected one connected vascular region with minimal fluorescence in the average projection of the data and calculated an initial vascular reference curve. This region was assumed to have the lowest possibility to contain any true GRAB_NE signal.
2.
To avoid compensating for slow changes in the true GRAB_NE signal, we subtracted the curve of each pixel by a 100-frame moving average.
3.
We applied linear regression and fit the logarithm of the processed curve to the initial vascular reference curve. The exponential of the fit data was then taken to represent the hemodynamic effect for each pixel. To account for cases where the initial reference curve was contaminated, we iteratively refined the reference curve before fitting, that is, we calculated the weighted average of the original curve for all pixels (fitting parameter a in $e^{ah_k\left( t \right) + b}$ for each pixel is considered as the weight) and subtracted its moving average.
4.
We removed the hemodynamic artifact (if any) by dividing the raw pixel curve by the exponential of the fitting data, $F_k\left( t \right) \approx \frac{{Y_k\left( t \right)}}{{e^{h_k(t)}}}$.

Next, only the least (1–25%) hemodynamically affected pixels with the lowest a were taken, excluding the bottom 1%, which often contained artifacts, and these were averaged together and used as the final GRAB_NE signal. For spectral analysis, each recording was concatenated together in 10 min segments, padding with its median value if necessary, and then run in Chronux with no frequency padding, a time bandwidth of 3, five tapers, and passing frequencies above half the window size (3 × 10⁻³ Hz) and less than the Nyquist frequency (0.9 Hz). To identify phasic increases in the GRAB_NE signal, built-in MATLAB functions were used to determine local peaks in the signal and a prominence threshold was used to determine the different magnitudes of GRAB_NE increases.

Event-triggered averages

All averages were computed by identifying events (for example, movement offset, pupil dilation and so on) and taking data in a symmetric time window around the events. The data are subsequently plotted as the mean and standard error across all events, except for spectrograms where the median was used.

Correlations

For comparisons of maximum and change in astrocyte Ca²⁺/pupil/speed after arousal, values were computed for each trace separately and then linearly correlated with P values describing the probability of a true R² relationship between each two metrics. For correlations between astrocyte and neuronal Ca²⁺ activity, the average population fluorescence was taken for each and z-scored before cross-correlation. For behavioral state-separated cross-correlations, the same procedure applied, but only z-scored data from either moving periods or stationary periods were used. For correlations within neuronal and astrocyte populations, the pairwise correlation between each cell (neurons) or event (astrocytes) was computed, the symmetric and autocorrelations were excluded, and the overall mean was taken to obtain a single value indicating the synchrony of Ca²⁺ dynamics within each cellular population. For cross-correlations between pupil and imaging data (GRAB_NE and astrocyte Ca²⁺), all data were z-scored, resampled to either 10 or 30 Hz, padded with nan values if unequal in length, and then cross-correlated. For cross-correlations between LFP band power and Ca²⁺ imaging data, all data were z-scored, averaged and then resampled to 2 Hz to match the LFP resolution before cross-correlation.

hSYN-hM4Di

To estimate the effect of CNO on astrocyte Ca²⁺, the average Ca²⁺ event properties and overall event rate for each recording were randomly sampled 10⁴ times and CNO data were subtracted from corresponding baseline data. This procedure generated the range of treatment effects possible from the sampled data, and a P value was calculated as the proportion of CNO difference from baseline that was less than the maximum, or greater than the minimum, difference found in saline conditions. For calculating the modulation of astrocyte Ca²⁺ responses to arousal, the absolute change in average astrocyte Ca²⁺ fluorescence after either movement onset or pupil dilation was determined. Outliers during small stationary responses, which might reflect the influence of other variables on astrocyte Ca²⁺, were excluded. The magnitude of astrocyte Ca²⁺ responses to arousal after treatment was then compared with baseline responses.

Neuronal arousal PC analysis

PCA was done using the built-in MATLAB function on z-scored neuronal Ca²⁺ data. The pupil diameter and PC data were then resampled to an effective rate of 10 Hz, and the Pearson’s correlation between the pupil diameter and each PC was used to identify the arousal PC for each recording This PC was then normalized to the maximum value before subsequent analysis.

For comparisons between wild-type and Adra1a^fl/fl mice, the response to each movement onset or stationary pupil dilation was normalized to the median value in the window, and then the average arousal PC value during the event period was taken with a two-frame offset to account for a slight lag in the arousal-associated neuronal response.

Fiber photometry

Photometry data were detrended using linear regression to correct bleaching and normalized by z-scoring. For startle experiments, recordings were denoised using a FIR filter (cutoff at 2 Hz, transition width 0.5 Hz). Startle responses were aligned to the onset of the jRGECO signal by fitting a sigmoid to the evoked jRGECO and taking the fourth derivative to identify the onset inflection point.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The data presented in this study are publicly available on Dryad (https://doi.org/10.7272/Q6XK8CS6).

Code availability

The code used to generate the findings of this study is publicly available on Zenodo (https://doi.org/10.5281/zenodo.7098082).

Change history

03 May 2023
In the version of this article initially published, the legends for Supplementary Videos 3 and 4 were interchanged and have now been updated in the HTML version of the article.

References

Harris, K. D. & Thiele, A. Cortical state and attention. Nat. Rev. Neurosci. 12, 509–523 (2011).
Article CAS PubMed PubMed Central Google Scholar
Crochet, S. & Petersen, C. C. H. Correlating whisker behavior with membrane potential in barrel cortex of awake mice. Nat. Neurosci. 9, 608–610 (2006).
Article CAS PubMed Google Scholar
Niell, C. M. & Stryker, M. P. Modulation of visual responses by behavioral state in mouse visual cortex. Neuron 65, 472–479 (2010).
Article CAS PubMed PubMed Central Google Scholar
Steriade, M., McCormick, D. A. & Sejnowski, T. J. Thalamocortical oscillations in the sleeping and aroused brain. Science 262, 679–685 (1993).
Article CAS PubMed Google Scholar
Fries, P. Neuronal gamma-band synchronization as a fundamental process in cortical computation. Annu. Rev. Neurosci. 32, 209–224 (2009).
Article CAS PubMed Google Scholar
Vinck, M., Batista-Brito, R., Knoblich, U. & Cardin, J. A. Arousal and locomotion make distinct contributions to cortical activity patterns and visual encoding. Neuron 86, 740–754 (2015).
Article CAS PubMed PubMed Central Google Scholar
McGinley, M. J. et al. Waking state: rapid variations modulate neural and behavioral responses. Neuron 87, 1143–1161 (2015).
Article CAS PubMed PubMed Central Google Scholar
Einstein, M. C., Polack, P.-O., Tran, D. T. & Golshani, P. Visually evoked 3–5 Hz membrane potential oscillations reduce the responsiveness of visual cortex neurons in awake behaving mice. J. Neurosci. 37, 5084–5098 (2017).
Article CAS PubMed PubMed Central Google Scholar
Liu, Y., Rodenkirch, C., Moskowitz, N., Schriver, B. & Wang, Q. Dynamic lateralization of pupil dilation evoked by locus coeruleus activation results from sympathetic, not parasympathetic, contributions. Cell Rep. 20, 3099–3112 (2017).
Article CAS PubMed PubMed Central Google Scholar
Polack, P.-O., Friedman, J. & Golshani, P. Cellular mechanisms of brain state-dependent gain modulation in visual cortex. Nat. Neurosci. 16, 1331–1339 (2013).
Article CAS PubMed PubMed Central Google Scholar
Constantinople, C. M. & Bruno, R. M. Effects and mechanisms of wakefulness on local cortical networks. Neuron 69, 1061–1068 (2011).
Article CAS PubMed PubMed Central Google Scholar
Nakadate, K., Imamura, K. & Watanabe, Y. Cellular and subcellular localization of alpha-1 adrenoceptors in the rat visual cortex. Neuroscience 141, 1783–1792 (2006).
Article CAS PubMed Google Scholar
Liu, Y. U. et al. Neuronal network activity controls microglial process surveillance in awake mice via norepinephrine signaling. Nat. Neurosci. 22, 1771–1781 (2019).
Article CAS PubMed PubMed Central Google Scholar
Ding, F. et al. α1-Adrenergic receptors mediate coordinated Ca²⁺ signaling of cortical astrocytes in awake, behaving mice. Cell Calcium 54, 387–394 (2013).
Article CAS PubMed Google Scholar
Oe, Y. et al. Distinct temporal integration of noradrenaline signaling by astrocytic second messengers during vigilance. Nat. Commun. 11, 471 (2020).
Article CAS PubMed PubMed Central Google Scholar
Paukert, M. et al. Norepinephrine controls astroglial responsiveness to local circuit activity. Neuron 82, 1263–1270 (2014).
Article CAS PubMed PubMed Central Google Scholar
Ye, L. et al. Ethanol abolishes vigilance-dependent astroglia network activation in mice by inhibiting norepinephrine release. Nat. Commun. 11, 6157 (2020).
Article CAS PubMed PubMed Central Google Scholar
Ma, Z., Stork, T., Bergles, D. E. & Freeman, M. R. Neuromodulators signal through astrocytes to alter neural circuit activity and behaviour. Nature 539, 428–432 (2016).
Article CAS PubMed PubMed Central Google Scholar
Mu, Y. et al. Glia accumulate evidence that actions are futile and suppress unsuccessful behavior. Cell 178, 27–43.e19 (2019).
Article CAS PubMed Google Scholar
Bojarskaite, L. et al. Astrocytic Ca²⁺ signaling is reduced during sleep and is involved in the regulation of slow wave sleep. Nat. Commun. 11, 3240 (2020).
Article CAS PubMed PubMed Central Google Scholar
Poskanzer, K. E. & Yuste, R. Astrocytes regulate cortical state switching in vivo. Proc. Natl Acad. Sci. USA 113, E2675–E2684 (2016).
Article CAS PubMed PubMed Central Google Scholar
Wang, Y. et al. Accurate quantification of astrocyte and neurotransmitter fluorescence dynamics for single-cell and population-level physiology. Nat. Neurosci. 22, 1936–1944 (2019).
Article CAS PubMed PubMed Central Google Scholar
McGinley, M. J., David, S. V. & McCormick, D. A. Cortical membrane potential signature of optimal states for sensory signal detection. Neuron 87, 179–192 (2015).
Article CAS PubMed PubMed Central Google Scholar
Reimer, J. et al. Pupil fluctuations track fast switching of cortical states during quiet wakefulness. Neuron 84, 355–362 (2014).
Article CAS PubMed PubMed Central Google Scholar
Aston-Jones, G. & Cohen, J. D. Adaptive gain and the role of the locus coeruleus-norepinephrine system in optimal performance. J. Comp. Neurol. 493, 99–110 (2005).
Article CAS PubMed Google Scholar
Feng, J. et al. A genetically encoded fluorescent sensor for rapid and specific in vivo detection of norepinephrine. Neuron 102, 745–761.e8 (2019).
Article CAS PubMed PubMed Central Google Scholar
Ma, Y. et al. Wide-field optical mapping of neural activity and brain haemodynamics: considerations and novel approaches. Philos. Trans. R. Soc. Lond. B 371, 20150360 (2016).
Article Google Scholar
Reimer, J. et al. Pupil fluctuations track rapid changes in adrenergic and cholinergic activity in cortex. Nat. Commun. 7, 13289 (2016).
Article CAS PubMed PubMed Central Google Scholar
Stringer, C. et al. Spontaneous behaviors drive multidimensional, brainwide activity. Science 364, 255 (2019).
Article PubMed PubMed Central Google Scholar
Armbruster, B. N., Li, X., Pausch, M. H., Herlitze, S. & Roth, B. L. Evolving the lock to fit the key to create a family of G protein-coupled receptors potently activated by an inert ligand. Proc. Natl Acad. Sci. USA 104, 5163–5168 (2007).
Article PubMed PubMed Central Google Scholar
Herreras, O. Local field potentials: myths and misunderstandings. Front. Neural Circuits 10, 101 (2016).
Article PubMed PubMed Central Google Scholar
Pankratov, Y. & Lalo, U. Role for astroglial α1-adrenoreceptors in gliotransmission and control of synaptic plasticity in the neocortex. Front. Cell Neurosci. 9, 230 (2015).
Article PubMed PubMed Central Google Scholar
Farhy-Tselnicker, I. et al. Activity-dependent modulation of synapse-regulating genes in astrocytes. eLife 10, e70514 (2021).
Article CAS PubMed PubMed Central Google Scholar
Bayraktar, O. A. et al. Astrocyte layers in the mammalian cerebral cortex revealed by a single-cell in situ transcriptomic map. Nat. Neurosci. 23, 500–509 (2020).
Article CAS PubMed PubMed Central Google Scholar
Speed, A., Del Rosario, J., Burgess, C. P. & Haider, B. Cortical state fluctuations across layers of V1 during visual spatial perception. Cell Rep. 26, 2868–2874.e3 (2019).
Article CAS PubMed PubMed Central Google Scholar
Wahis, J. & Holt, M. G. Astrocytes, Noradrenaline, α1-adrenoreceptors, and neuromodulation: evidence and unanswered questions. Front. Cell Neurosci. 15, 645691 (2021).
Article CAS PubMed PubMed Central Google Scholar
Papouin, T., Dunphy, J. M., Tolman, M., Dineley, K. T. & Haydon, P. G. Septal cholinergic neuromodulation tunes the astrocyte-dependent gating of hippocampal NMDA receptors to wakefulness. Neuron 94, 840–854.e7 (2017).
Article CAS PubMed PubMed Central Google Scholar
Di Castro, M. A. et al. Local Ca²⁺ detection and modulation of synaptic release by astrocytes. Nat. Neurosci. 14, 1276–1284 (2011).
Article PubMed Google Scholar
Bernardinelli, Y. et al. Activity-dependent structural plasticity of perisynaptic astrocytic domains promotes excitatory synapse stability. Curr. Biol. 24, 1679–1688 (2014).
Article CAS PubMed Google Scholar
Stumpf, C. & Pichler, L. Synchronizing effect of the alpha 1-adrenoceptor agonist 2-(2-chloro-5-trifluoromethylphenylimino)imidazolidine on rabbit electroencephalogram. Arzneimittelforschung 38, 770–774 (1988).
CAS PubMed Google Scholar
Berridge, C. W. & Spencer, R. C. Differential cognitive actions of norepinephrine a2 and a1 receptor signaling in the prefrontal cortex. Brain Res. 1641, 189–196 (2016).
Article CAS PubMed Google Scholar
Perez, D. M. α1-Adrenergic receptors in neurotransmission, synaptic plasticity, and cognition. Front. Pharmacol. 11, 581098 (2020).
Article CAS PubMed PubMed Central Google Scholar
Thiele, A. & Bellgrove, M. A. Neuromodulation of attention. Neuron 97, 769–785 (2018).
Article CAS PubMed PubMed Central Google Scholar
Shimaoka, D., Harris, K. D. & Carandini, M. Effects of arousal on mouse sensory cortex depend on modality. Cell Rep. 22, 3160–3167 (2018).
Article CAS PubMed PubMed Central Google Scholar
Lee, A. M. et al. Identification of a brainstem circuit regulating visual cortical state in parallel with locomotion. Neuron 83, 455–466 (2014).
Article CAS PubMed PubMed Central Google Scholar
Celada, P., Puig, M. V. & Artigas, F. Serotonin modulation of cortical neurons and networks. Front Integr. Neurosci. 7, 25 (2013).
Article CAS PubMed PubMed Central Google Scholar
Speed, A., Del Rosario, J., Mikail, N. & Haider, B. Spatial attention enhances network, cellular and subthreshold responses in mouse visual cortex. Nat. Commun. 11, 505 (2020).
Article CAS PubMed PubMed Central Google Scholar
Slezak, M. et al. Distinct mechanisms for visual and motor-related astrocyte responses in mouse visual cortex. Curr. Biol. 29, 3120–3127.e5 (2019).
Article CAS PubMed PubMed Central Google Scholar
Stobart, J. L. et al. Cortical circuit activity evokes rapid astrocyte calcium signals on a similar timescale to neurons. Neuron 98, 726–735.e4 (2018).
Article CAS PubMed Google Scholar
Lines, J., Martin, E. D., Kofuji, P., Aguilar, J. & Araque, A. Astrocytes modulate sensory-evoked neuronal network activity. Nat. Commun. 11, 3689 (2020).
Article CAS PubMed PubMed Central Google Scholar
Batiuk, M. Y. et al. Identification of region-specific astrocyte subtypes at single cell resolution. Nat. Commun. 11, 1220 (2020).
Article CAS PubMed PubMed Central Google Scholar
Johnson, M. B. & Walsh, C. A. Cerebral cortical neuron diversity and development at single-cell resolution. Curr. Opin. Neurobiol. 42, 9–16 (2017).
Article PubMed Google Scholar
Pala, A. & Petersen, C. C. State-dependent cell-type-specific membrane potential dynamics and unitary synaptic inputs in awake mice. eLife 7, e35869 (2018).
Article PubMed PubMed Central Google Scholar
Deemyad, T., Lüthi, J. & Spruston, N. Astrocytes integrate and drive action potential firing in inhibitory subnetworks. Nat. Commun. 9, 4336 (2018).
Article PubMed PubMed Central Google Scholar
Zucca, S. et al. An inhibitory gate for state transition in cortex. eLife 6, e26177 (2017).
Article PubMed PubMed Central Google Scholar
Rasmussen, R. et al. Cortex-wide changes in extracellular potassium ions parallel brain state transitions in awake behaving mice. Cell Rep. 28, 1182–1194.e4 (2019).
Article CAS PubMed PubMed Central Google Scholar
Tewari, B. P. et al. Perineuronal nets decrease membrane capacitance of peritumoral fast spiking interneurons in a model of epilepsy. Nat. Commun. 9, 4724 (2018).
Article PubMed PubMed Central Google Scholar
Ding, F. et al. Changes in the composition of brain interstitial ions control the sleep–wake cycle. Science 352, 550–555 (2016).
Article CAS PubMed PubMed Central Google Scholar
Verkhratsky, A. Physiology of neuronal–glial networking. Neurochem. Int. 57, 332–343 (2010).
Article CAS PubMed Google Scholar
Mazaud, D., Capano, A. & Rouach, N. The many ways astroglial connexins regulate neurotransmission and behavior. Glia 69, 2527–2545 (2021).
Article PubMed Google Scholar
Goldey, G. J. et al. Removable cranial windows for long-term imaging in awake mice. Nat. Protoc. 9, 2515–2538 (2014).
Article CAS PubMed PubMed Central Google Scholar
Watson, B. O., Yuste, R. & Packer, A. M. PackIO and EphysViewer: software tools for acquisition and analysis of neuroscience data. Preprint at bioRxiv https://doi.org/10.1101/054080 (2016).
Saravanan, V., Berman, G. J. & Sober, S. J. Application of the hierarchical bootstrap to multi-level data in neuroscience. Neuron. Behav. Data Anal. Theory 3, (2020).
Bokil, H., Andrews, P., Kulkarni, J. E., Mehta, S. & Mitra, P. P. Chronux: a platform for analyzing neural signals. J. Neurosci. Methods 192, 146–151 (2010).
Article PubMed PubMed Central Google Scholar
Pachitariu, M. et al. Suite2p: beyond 10,000 neurons with standard two-photon microscopy. Preprint at bioRxiv https://doi.org/10.1101/061507 (2016).
Watson, B. O., Levenstein, D., Greene, J. P., Gelinas, J. N. & Buzsáki, G. Network homeostasis and state dynamics of neocortical sleep. Neuron 90, 839–852 (2016).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank G. Chin and S. Yokoyama for technical assistance, K. Roberts and T. Li for help with smFISH data generation, J. Thompson for administrative assistance, and R. Reitman for assistance in data analysis pipeline construction. We also thank the Poskanzer lab for helpful discussions about the project, and S. Lavrentyeva and E. Kish for comments on the paper. We thank E. Feinberg and C. Kirst for helpful discussions on the resubmitted paper and experiments, and J. Reimer and C. Smith for discussions related to neuron-astrocyte imaging. This work was funded by the UCSF Genentech Fellowship (M.E.R.); National Institute of Health R01MH110504 (G.Y.), U19NS123719 (G.Y.), R01NS099254 (K.E.P.), R01MH121446 (K.E.P.) and R01HL31113-27 (P.C.S.); National Science Foundation 1750931 (G.Y.) and CAREER 1942360 (K.E.P.); Veteran Affairs BX004314 (P.C.S.); and the UCSF Program for Breakthrough Biomedical Research, which is funded in part by the Sandler Foundation (K.E.P.). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the paper.

Author information

Authors and Affiliations

Neuroscience Graduate Program, University of California, San Francisco, San Francisco, CA, USA
Michael E. Reitman, Drew D. Willoughby & Kira E. Poskanzer
Department of Biochemistry & Biophysics, University of California, San Francisco, San Francisco, CA, USA
Michael E. Reitman, Vincent Tse, Drew D. Willoughby, Alba Peinado & Kira E. Poskanzer
Bradley Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University, Arlington, VA, USA
Xuelong Mi & Guoqiang Yu
Wellcome Sanger Institute, Cambridge, UK
Alexander Aivazidis & Omer A. Bayraktar
Department of Medicine and Research Service, San Francisco Veterans Affairs Medical Center and Cardiovascular Research Institute, University of California, San Francisco, San Francisco, CA, USA
Bat-Erdene Myagmar & Paul C. Simpson
Kavli Institute for Fundamental Neuroscience, San Francisco, CA, USA
Kira E. Poskanzer

Authors

Michael E. Reitman
View author publications
You can also search for this author in PubMed Google Scholar
Vincent Tse
View author publications
You can also search for this author in PubMed Google Scholar
Xuelong Mi
View author publications
You can also search for this author in PubMed Google Scholar
Drew D. Willoughby
View author publications
You can also search for this author in PubMed Google Scholar
Alba Peinado
View author publications
You can also search for this author in PubMed Google Scholar
Alexander Aivazidis
View author publications
You can also search for this author in PubMed Google Scholar
Bat-Erdene Myagmar
View author publications
You can also search for this author in PubMed Google Scholar
Paul C. Simpson
View author publications
You can also search for this author in PubMed Google Scholar
Omer A. Bayraktar
View author publications
You can also search for this author in PubMed Google Scholar
Guoqiang Yu
View author publications
You can also search for this author in PubMed Google Scholar
Kira E. Poskanzer
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

M.E.R. and K.E.P. conceptualized and designed the experiments. M.E.R. performed all 2P imaging and physiology experiments, and carried out data analyses for these experiments. V.T. performed immunostaining experiments and analyses. X.M. and G.Y. worked with M.E.R. and K.E.P. to develop the GRAB_NE analysis method. D.D.W. performed and analyzed the fiber photometry experiments. A.P. collaborated on the analysis of 2P imaging and physiology data. A.A. and O.A.B. performed and analyzed smFISH experiments. B.-E.M. and P.C.S. generated Adra1a floxed mice. M.E.R. and K.E.P wrote the paper with input from other authors. K.E.P. supervised all phases of the project.

Corresponding author

Correspondence to Kira E. Poskanzer.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Neuroscience thanks Inbal Goshen, Nathan Smith and the other, anonymous reviewer(s) for their contribution to the peer review of this work. This article has been peer reviewed as part of Springer Nature’s Guided Open Access initiative.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Dissection of astrocyte Ca²⁺ and behavioral state.

(a) Representative 2P mean projection image from one ten-minute recording of in vivo astrocyte GcaMP6f (left, scale bar = 100 µm). AQuA detected both large (middle) and small (right) astrocyte Ca²⁺ events within the entire movie, even when they were spatially overlapping. Arrows indicate two pairs of spatially overlapping events (arrows 1 and 3, and arrows 2 and 4).(b) Traces from the AQuA events shown in (a), with the time period of the AQuA-detected event highlighted in red. (c) Related to Fig. 1c: Individual astrocyte Ca²⁺ events correlated better with pupil diameter (left, n = 1.2e⁴ Ca²⁺ events, One-sided Kruskal-Wallis test) and had a shorter lag with pupil diameter than wheel speed (right, pupil n = 9.6e³, wheel n = 8.3e³, rank-sum test). (d) Related to Fig. 1d: average pupil diameter during movement (n = 100) and stationary periods (n = 76). (e) Classification of behavioral state by both pupil diameter and movement. (f) Pupil dilation is smaller (left) and shorter (right, rank-sum tests) during stationary periods (n = 261 dilations, blue) compared with movement-associated dilations (n = 136 dilations, red, boxplots show median and IQR with whiskers to 1.5 * IQR, two-sided Rank Sum test). (g) Related to Fig. 1l: Left: Movement duration (n = 104) was related to the maximum wheel speed (top) and pupil diameter (middle), but not to the maximum astrocyte Ca²⁺ (bottom). Right: the latency to the maximum wheel speed (top), pupil diameter (middle), and astrocyte Ca²⁺ (bottom) were strongly linked to movement duration. (h) Heatmap summary of the r² between the latencies in (g) right, and movement bout duration.

Extended Data Fig. 2 Hemodynamic correction of 2P GRAB_NE signals.

(a) Left: Average projection of a single GRAB_NE recording with four regions-of-interest (ROI) in areas with clear GRAB_NE fluorescence (black squares, labeled 1–4) and one ROI in a blood vessel (red square, labeled BV). Middle: standard deviation projection. Right: average fluorescence in each ROI before hemodynamic correction. (b) Left: Estimated hemodynamic signal present in each pixel. Middle: recovered average projection after hemodynamic correction. Note average fluorescence after hemodynamic correction is broadly similar the uncorrected data in (a) with some blurring due to Gaussian smoothing in preprocessing. Right: average ROI fluorescence after hemodynamic correction. (c) ROI-free methodology for obtaining GRAB_NE signal. Left: Following hemodynamic correction, the bottom quartile of pixels with the lowest hemodynamic weights, excluding the bottom 1% which often had artefactual signals, were kept. Right: These pixels were averaged together to produce the corrected GRAB_NE signal (bottom) which had substantially less hemodynamic contamination than the uncorrected field fluorescence (top). (d) Correlation between the corrected GRAB_NE signal and the uncorrected ROIs in (a) and (b). (e) Left: Correlation between uncorrected GRAB_NE fluorescence and pupil diameter. Right: Example traces from the average of pixels that were positively correlated (top), negatively correlated (bottom), or showed little relationship (middle) with pupil diameter. (f) The corrected GRAB_NE signal is not similar to the uncorrected GRAB_NE signal (red dot) but instead reflects the GRAB_NE signal of binned pixels (black dots) correlated to pupil diameter (R² = 0.62, p = 1.55⁻⁵, two-sided t-test).

Extended Data Fig. 3 Freely moving fiber photometry recordings of GRAB_NE and astrocyte Ca²⁺.

(a) Schematic of fiber photometry recording set-up. (b) A 5-minute example fiber photometry recording of GRAB_NE signal. (c) GRAB_NE power spectrum. n = 4 mice, 19 recordings over multiple days, 10–30 minutes/recording. Error lines are ± s.e.m. across all mice.) (d) GRAB_NE (green) and astrocyte jRGECO1b (magenta) signals evoked by startle responses due to tail lifts (grey arrow heads). (e) Average startle-response evoked traces for GRAB_NE and jRGECO, aligned to onset of astrocyte Ca²⁺. n = 4 mice, 4 tail lifts per mouse. Error is ± s.e.m. across mice.

Extended Data Fig. 4 Analysis of contributions to astrocyte Ca²⁺.

(a) Left: A Random Forest Regression model was trained (80% of data, 2.3e⁴ samples) to predict (20% of data, 5.6³ samples) average astrocyte Ca²⁺ fluorescence accurately (mean r² = 0.85 ± 4.5e^-3 std, n = 10 cross-validations). Right: Relative feature importance based on permutation testing of each predictor. Randomly generated data was included as a negative control and did not inform model predictions. (b) Same analysis as in (a), using the neuronal arousal PC rather than average neuronal activity. Random Forest Regression using the neuronal arousal PC showed a similar accuracy to that using mean neuronal fluorescence (mean r² = 0.87 ± 6.4e^-3 std, n = 10 cross-validations). Boxplots show median and IQR with whiskers to 1.5 * IQR.

Extended Data Fig. 5 Cortical astrocyte expression of adrenergic receptors.

(a) Ribosomal-mRNA expression in visual cortex astrocytes of P120 mice from the Farhy-Tselnicker et. al. publicly available dataset. Visual cortex astrocytes show expression of many adrenergic receptors (left) but preferentially express the Adra1a receptor relative to input control (right). Boxplots show median and IQR with whiskers to 1.5 * IQR (N = 3 mice for astrocyte ribosomal-mRNA and 1 input control). (b) Left: Representative brain section with Adra1a (green), Adra1b (red), and Adra2a (yellow) mRNA labeled by smFISH using the LaST map pipeline³⁴. SLC1a3 (GLAST) mRNA expression (purple) was used to define the location of astrocytes. The white square shows a representative ROI used to quantify visual cortex expression. Right: High-magnification image of visual cortex, showing receptor mRNA both within astrocytes marked by high GLAST mRNA (purple), and outside of astrocytes. (c) Quantification of NE-receptor mRNA expression in visual cortex astrocytes, quantified by relative cortical depth. Line plots show mean ± std (n = 3 mice, 7 ROIs).

Extended Data Fig. 6 Generation of Adra1a^fl/fl mice.

Top: schematic of the genetic strategy used to generate conditional deletion of Adra1a. Primers (green arrows) were designed to identify and distinguish between the endogenous and knock-in Adra1a alleles. Bottom: Numbers are in base pairs. Validation of the knock-in strategy used to generate Adra1a^fl/fl mice. PCR of Adra1a^fl/+ mice generates both a 154 base pair band corresponding to the wild type allele and a 235 base pair band corresponding to successful LoxP knock-in, which is not found in wild-type mice. N = 5 Adra1a^fl/+ and 4 wild type mice.

Extended Data Fig. 7 Astrocyte-specific Cre expression.

(a) Representative average 2 P image of in vivo neuronal jRGECO1b (gray) and astrocytes expressing Cre-GFP (magenta) (b) Example images show astrocyte-specific Cre-GFP (magenta) expression overlaps with the astrocyte marker S100β (left, green) but not the neuronal marker NeuN (right, grey), scale bar = 100 µm. (c) Representative mean-projection of a confocal 60x z-stack used for cell counting. Each animal had two sections with each section having six distinct field-of-views containing 20 z-planes. GFAP(0.7)-Cre-GFP expression (left) was highly astrocytic and the neuronal marker NeuN (middle) was rarely colocalized (right, scale bars = 50 µm). (d) Cre expression in neurons (grey bars) was low across genotypes (overall = 2.2 ± 0.39 neurons/50 µm³) and no difference was found between Cre expression in Adra1a^fl/fl mice and control mice or the cohort as a whole (One-sided Kruskal-Wallis test). (e) Cre-expressing neurons were rare in both wild-type (black, 4.0% ± 0.75%) and Adra1a^fl/fl mice (green, 4.1% ± 0.65%, two-sided ranked-sum test). (n = 7 Adra1a^fl/fl and n = 5 wild-type mice for all graphs. Data are presented as mean ± S.E.M.).

Extended Data Fig. 8 Resonant galvo imaging (7.5 Hz effective frame rate) confirms increased neuronal activity in Adra1a^fl/fl mice.

(a–c) Histograms show neuronal activity in Adra1a^fl/fl (green) and wild-type littermate mice (grey) based on HB. (a) Neuronal Ca²⁺ event rate was increased in Adra1a^fl/fl (green, n = 4 mice) compared to wild-type mice (grey, n = 4 mice). (b) Left: power spectrum of Adra1a^fl/fl (green) and wild-type (black) Ca²⁺ activity between 0–0.05 Hz. Right: Total power in 0–0.05 Hz neuronal Ca²⁺ activity, comparing Adra1a^fl/fl and wild-type mice. (c) Same analysis as (b), comparing power in neuronal Ca²⁺ fluctuations > 0.05 Hz. All p-values from one-sided HB tests.

Extended Data Fig. 9 Quantification of arousal-associated neuronal activity.

(a) Cumulative distribution plots of the duration of movement bouts (left) and stationary pupil dilations (right) for wild-type (black, n = 400 movement bouts and 1396 pupil dilations) and Adra1a^fl/fl (green, n = 177 movement bouts and n = 923 pupil dilations) mice. (b) Analysis range (red) used to quantify the normalized arousal PC response to movement in wild-type (left) and Adra1a^fl/fl (right) mice. Each row represents one movement event, and the movements from all recordings and mice are concatenated to show the entire dataset (n = 4 wild-type mice and n = 4 Adra1a^fl/fl mice). (c) Same analysis as in (b), showing stationary pupil dilations in Adra1a^fl/fl mice compared to wild-type.

Supplementary information

Supplementary Information

Supplementary Tables 1–5.

Reporting Summary

2P imaging of GRAB_NE and astrocyte jRGECO1b. Thirty-minute recording of spontaneous GRAB_NE (green) and astrocyte jRGECO1b (magenta) fluctuations in layer 2/3 of visual cortex of an awake mouse. Recording was acquired with a ~410 µm × 410 µm field of view, 10× averaged, and a one-pixel median filter was applied. Video is played at a 10 Hz frame rate.

2P imaging of neuronal GCaMP6f and astrocyte jRGECO1b. Ten-minute recording of spontaneous neuronal GCaMP6f (gray) and astrocyte jRGECO1b (magenta) fluctuations in layer 2/3 of the awake cortex. Recording was acquired with a ~410 µm × 410 µm field of view, 10× averaged, and a 0.5-pixel median filter was applied. Video is played at a 10 Hz frame rate.

Neuronal and astrocyte Ca²⁺ responses to A61603. Neuronal GCaMP6f (gray, left) and astrocyte jRGECO1b (magenta, right) recorded for 30 min at baseline and for 40 min post-injection of A61603 (10 µg kg⁻¹, i.p.) Labels on the top left indicate pharmacological condition. Recording was acquired with a ~410 µm × 410 µm field of view, 10× averaged, and a 0.5-pixel median filter was applied. Video is played at a 60 Hz frame rate.

2P z-stack showing neuronal hM4Di and astrocyte GCaMP6f expression in same tissue region. An ~300 µm z-stack of in vivo expression of neuronal hM4Di-mCherry (gray) and astrocyte jRGECO1b (magenta). The z-stack was acquired with a 5 µm step size. A four-frame moving average projection (astrocytes) or s.d. projection (neurons) was calculated for a display step size of 20 µm per frame. Neuronal data were background-subtracted to aid visualization of soma, and astrocyte data were attenuation-corrected.

Supplementary Data

Editorial assessment report.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Reitman, M.E., Tse, V., Mi, X. et al. Norepinephrine links astrocytic activity to regulation of cortical state. Nat Neurosci 26, 579–593 (2023). https://doi.org/10.1038/s41593-023-01284-w

Download citation

Received: 04 May 2021
Accepted: 14 February 2023
Published: 30 March 2023
Issue Date: April 2023
DOI: https://doi.org/10.1038/s41593-023-01284-w

This article is cited by

Serotonergic modulation of vigilance states in zebrafish and mice
- Yang Zhao
- Chun-Xiao Huang
- Jianren Song
Nature Communications (2024)
Microglia regulate sleep through calcium-dependent modulation of norepinephrine transmission
- Chenyan Ma
- Bing Li
- Yang Dan
Nature Neuroscience (2024)
A midbrain GABAergic circuit constrains wakefulness in a mouse model of stress
- Shuancheng Ren
- Cai Zhang
- Zhian Hu
Nature Communications (2024)
Multicore fiber optic imaging reveals that astrocyte calcium activity in the mouse cerebral cortex is modulated by internal motivational state
- Yung-Tian A. Gau
- Eric T. Hsu
- Dwight E. Bergles
Nature Communications (2024)
Network-level encoding of local neurotransmitters in cortical astrocytes
- Michelle K. Cahill
- Max Collard
- Kira E. Poskanzer
Nature (2024)

Subjects

Abstract

Similar content being viewed by others

Main

Results

Astrocyte Ca2+ correlates with increases in pupil diameter

Phasic increases in NE precede astrocyte Ca2+

Astrocyte Ca2+ may reduce arousal-driven neuronal activity

Arousal-driven astrocyte Ca2+ is not local neuron-dependent

NE ties astrocyte Ca2+ to changes in cortical state

Enrichment of Adra1a mRNA in visual cortex astrocytes

Astrocyte Adra1a receptor signaling shapes neuronal activity

NE signaling via astrocyte Adra1a modulates cortical state

Discussion

Astrocytes regulate arousal-associated cortical circuits

Astrocyte Ca2+ is poised to regulate neuronal activity

Receptor-specific relationship between NE and cortical state

What is arousal?

How does arousal modulate different cellular subpopulations?

Methods

Animals

Generation of Adra1afl/fl mice

Surgical procedures and viral infection

Immunohistochemistry and in situ hybridization

Immunohistochemistry

Single-molecule fluorescent in situ hybridization

Recording setup

Pupillometry

2P imaging

Electrophysiology

In vivo pharmacology

Fiber photometry

Statistics and reproducibility

Speed calculations

Pupillometry

LFP

Ca2+ imaging

Machine-learning based analysis of input contribution to astrocyte Ca2+

GRABNE analysis

Event-triggered averages

Correlations

hSYN-hM4Di

Neuronal arousal PC analysis

Fiber photometry

Reporting summary

Data availability

Code availability

Change history

03 May 2023

References

Acknowledgements

Author information

Authors and Affiliations

Contributions

Corresponding author

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Extended data

Supplementary information

Rights and permissions

About this article

Cite this article

Share this article

This article is cited by

Search

Quick links

Astrocyte Ca²⁺ correlates with increases in pupil diameter

Phasic increases in NE precede astrocyte Ca²⁺

Astrocyte Ca²⁺ may reduce arousal-driven neuronal activity

Arousal-driven astrocyte Ca²⁺ is not local neuron-dependent

NE ties astrocyte Ca²⁺ to changes in cortical state

Astrocyte Ca²⁺ is poised to regulate neuronal activity

Generation of Adra1a^fl/fl mice

Ca²⁺ imaging

Machine-learning based analysis of input contribution to astrocyte Ca²⁺

GRAB_NE analysis