## Abstract

Neural activity regulates local increases in cerebral blood flow (ΔCBF) and the cortical metabolic rate of oxygen (ΔCMRO2) that constitutes the basis of BOLD functional neuroimaging signals. Glutamate signaling plays a key role in brain vascular and metabolic control; however, the modulatory effect of GABA is incompletely understood. Here we performed in vivo studies in mice to investigate how THIP (which tonically activates extrasynaptic GABAARs) and Zolpidem (a positive allosteric modulator of synaptic GABAARs) impact stimulation-induced ΔCBF, ΔCMRO2, local field potentials (LFPs), and fluorescent cytosolic Ca2+ transients in neurons and astrocytes. Low concentrations of THIP increased ΔCBF and ΔCMRO2 at low stimulation frequencies. These responses were coupled to increased synaptic activity as indicated by LFP responses, and to Ca2+ activities in neurons and astrocytes. Intermediate and high concentrations of THIP suppressed ΔCBF and ΔCMRO2 at high stimulation frequencies. Zolpidem had similar but less-pronounced effects, with similar dependence on drug concentration and stimulation frequency. Our present findings suggest that slight increases in both synaptic and extrasynaptic GABAAR activity might selectively gate and amplify transient low-frequency somatosensory inputs, filter out high-frequency inputs, and enhance vascular and metabolic responses that are likely to be reflected in BOLD functional neuroimaging signals.

## Introduction

A normal brain responds to increased neuronal activity with rises in local cerebral blood flow (ΔCBF) and in the metabolic rate of oxygen (ΔCMRO2) (Raichle and Mintun 2006). The tight coupling between neuronal activity and vascular supply and metabolism is regulated by neurotransmitter signaling, which controls ΔCBF mainly through calcium-dependent mechanisms in neurons and astrocytes (Lauritzen 2005; Iadecola and Nedergaard 2007; Koehler et al. 2009; Lind et al. 2013; Nizar et al. 2013), and ΔCMRO2 through ATP usage by different signaling components in cerebral gray matter (Erecinska and Silver 1989; Attwell and Laughlin 2001; Mathiesen et al. 2011). Stimulation-induced hemodynamic and metabolic responses produced by synaptic excitation are related to the effects of glutamate on neurons and astrocytes. However, despite growing evidence of the importance of cortical interneurons for neurovascular coupling, the specific role of GABA (γ-amino-butyric acid) receptors (GABAARs) remains unclear (Fergus and Lee 1997; Cauli et al. 2004; Niessing et al. 2005; Kocharyan et al. 2008; Enager et al. 2009; Cauli and Hamel 2010; Lecrux et al. 2011).

GABA is released from interneurons—which comprise 15–20% of the cortical neuronal population—and exerts its principal action by binding to low-affinity ligand-gated GABAARs on the postsynaptic membrane (Buzsaki et al. 2007). This transient and target-specific effect is known as “phasic” inhibition (Farrant and Nusser 2005). GABA from synaptic spillover and ambient GABA also diffuse and bind to high-affinity extrasynaptic GABAARs, giving rise to persistent “tonic” inhibition in the neocortex (Semyanov et al. 2004; Farrant and Nusser 2005; Drasbek et al. 2007). Furthermore, GABA can activate metabotropic G protein-coupled GABAB receptors in the CNS and ligand-gated GABAC receptors in the retina (Jacob et al. 2008). Compared with GABAARs, GABAB receptors elicit slower responses to GABA and reportedly induce hippocampal microvessel vasoconstriction in vitro (Fergus and Lee 1997). Previous studies in rodent cortex and cerebellum show that GABAAR activity modulates ΔCBF and ΔCMRO2 during increased cortical activity (Caesar et al. 2003; Caesar et al. 2008; Harris et al. 2010; Lecrux et al. 2011; Mathiesen et al. 2011), but it is unknown to what extent these responses are affected by synaptic versus extrasynaptic GABAARs.

The present study tested the hypothesis that increased activity of synaptic or extrasynaptic GABAARs differentially modulates the vascular and metabolic responses to glutamate-mediated rises in neuronal activity. Our results show that increased GABAAR activity bidirectionally affected local field potentials (LFPs), Ca2+ responses, ΔCBF, and ΔCMRO2 during somatosensory stimulation activity. We observed augmented responses at low drug concentrations and low stimulation frequencies and depressed responses at high stimulation frequencies and high drug doses, thus indicating that low-frequency activity in cortical neurons is particularly sensitive to small increments of GABAAR activity. Our findings demonstrated that glutamate–GABA interaction modulated stimulation-induced increases of ΔCBF and ΔCMRO2 and that slight rises in GABA enhanced stimulation-induced vascular and metabolic responses, and possibly also BOLD functional neuroimaging signals. Tonic inhibition had more pronounced effects on these responses than phasic inhibition. The observed disinhibition of excitatory neurons in the presence of slight GABAA increases may be due to the activation of interneuronal GABAergic autapses (THIP) or to increased interneuron-to-interneuron synaptic transmission (Zolpidem) (Hioki et al. 2013), whereas the depressant effect of increased tonic activity could be explained by decreased excitability of pyramidal cells (Drasbek and Jensen 2006).

## Materials and Methods

### Animal Handling

All procedures involving animals were approved by the Danish National Ethics Committee according to the guidelines set forth in the European Council's Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes. Twenty-eight 8-week-old male NMRI mice (Crl:NMRI(male)) were used. The trachea was cannulated for mechanical ventilation (SAAR-830; CWE), and catheters were placed into the left femoral artery and vein for infusion of substances and to monitor blood pressure and blood gasses. To ensure that the animals were kept at physiological conditions, we continuously monitored end-expiratory CO2 (microCapstar End-tidal CO2 Analyzer, CWE) and blood pressure (Pressure Monitor BP-1, World Precision Instruments) and repeatedly assessed blood gases in arterial blood samples (pO2, 95–110 mmHg; pCO2, 35–40 mmHg; pH, 7.35–7.45) using an ABL 700Series Radiometer. Body temperature was maintained at 37°C using a rectal temperature probe and heating blanket (Model TC-1000 Temperature Controller, CWE).

The experimental setup involved gluing the skull to a metal plate with cyanoacrylate gel (Loctite Adhesives). A craniotomy was drilled with a diameter of ∼4 mm, centered at 0.5 mm behind and 3 mm to the right of bregma over the sensory barrel cortex region. After dura removal, the preparation was covered with 0.75% agarose gel (type III-A, low EEO; Sigma–Aldrich), moistened with artificial cerebrospinal fluid (aCSF; NaCl 120 mm, KCl 2.8 mm, NaHCO3 22 mm, CaCl2 1.45 mm, Na2HPO4 1 mm, MgCl2 0.876 mm, and glucose 2.55 mm; pH = 7.4), and kept at 37°C and aerated with 95% air/5% CO2. For imaging experiments, part of the craniotomy was covered with a glass coverslip that permitted insertion of electrodes and pharmacological interventions. In 14 mice, we measured CBF, tissue partial pressure of oxygen (tpO2), and excitatory synaptic currents as indicated by extracellular LFPs in response to whisker pad stimulation. In the other 10 mice, we examined cytosolic Ca2+ responses using two-photon microscopy with the same stimulation protocol.

### Anesthesia

Anesthesia was induced with bolus injections of the α2-adrenergic receptor agonist xylazine (10 mg kg−1 i.p.) and the NMDA-receptor antagonist ketamine (60 mg kg−1 i.p.) and was maintained during surgery with supplemental doses of ketamine (30 mg kg−1/20 min i.p.). Upon completion of all surgical procedures, anesthesia was switched to continuous infusion (i.v.) with α-chloralose (0.01 mL 10 g−1 i.v.). At the end of the experimental protocol, the mice were euthanized by intravenous injection of air followed by decapitation.

Xylazine provides analgesia and sedation during surgical procedures but has substantial systemic effects on blood pressure and respiration. Ketamine does not affect the respiratory system but diminishes neurological responses due to NMDA-receptor antagonism (Hildebrandt et al. 2008). Both compounds are rapidly metabolized and fully excreted via the kidney and liver. On the other hand, the specific pharmacological effects of α-chloralose are not fully elucidated but likely involve the GABAergic inhibitory system (Garrett and Gan 1998). Alpha-chloralose is traditionally used within the field of neuroscience because it preserves neurovascular coupling better than other anesthetics, especially in relation to neuronal activity and CBF responses (Lindauer et al. 1993; Bonvento et al. 1994; Austin et al. 2005; Sumiyoshi et al. 2012).

### Whisker Pad Stimulation

The mouse sensory barrel cortex was activated by stimulation of the contralateral ramus infraorbitalis of the trigeminal nerve, using a set of custom-made bipolar electrodes inserted percutaneously. The cathode was positioned corresponding to the hiatus infraorbitalis, whereas the anode was inserted into the masticatory muscles (Nielsen and Lauritzen 2001).

Stimulation was administered for 1 ms with 1.5 mA and in trains of 15 s at 0.5, 1.0, 2.0, 3.0, and 5.0 Hz (ISO-flex, A.M.P.I., Israel) and was controlled by a sequencer file running within the Spike2 software (version 7.02, Cambridge Electronic Design, Cambridge, United Kingdom).

### Electrophysiology

Extracellular LFPs were recorded with a single-barreled glass microelectrode filled with isotonic NaCl (impedance, 2–3 MΩ; tip, 2 μm). The electrode was inserted into the whisker cortex (layers 2/3) whereas an Ag/AgCl ground electrode was placed near the cranial window submerged in aCSF. During the two-photon experiments, the Ag/AgCl ground electrode was inserted into the neck muscles. The signal was first amplified using a differential amplifier (gain, 10×; bandwidth, 0.1–10,000 Hz; DP-311 Warner Instruments), followed by additional amplification using CyberAmp 380 (gain, 100×; bandwidth, 0.1–10,000 Hz; Axon Instruments). Digital sampling was performed using the 1401 mkII interface (Cambridge Electronic Design) connected to Spike2 software (Cambridge Electronic Design) running at a sampling rate of 5 kHz. The LFPs were averaged for each stimulation train, and amplitudes were calculated as the difference between baseline and the first negative peak (excitatory postsynaptic potential; EPSP) (Creutzfeldt 1995).

We additionally calculated the summed field potential (ΣLFP)—that is, the product of the LFP amplitude (in mV) and the stimulus rate (in Hz)—as a summary measure of the extracellular ionic fluxes produced by synaptic activity reflecting Na+ entry via ionotropic glutamate receptors (Mathiesen et al. 2000). Thus, ΣLFP gives a measure of the summed synaptic activity during the entire stimulation train.

### Cortical Blood Flow Measurement

Cortical blood flow was monitored using laser Doppler flowmetry (LDF). An LDF probe (wavelength, 780 nm; 250-μm fiber separation allowing CBF measurement up to 1 mm of deep; Perimed, PeriFlux 4001 Master, Sweden) that continuously recorded CBF was placed at a fixed position 0.3 mm above the pial surface in a region devoid of large vessels (Fabricius et al. 1997) near the microelectrodes recording electrophysiological variables and oxygen tension (∼0.1 mm apart). The LDF signal was smoothed with a time constant of 0.2 s, sampled at 10 Hz, A/D-converted, digitally recorded, and smoothed again with a time constant of 1 s using the Spike2 software. The LDF method does not measure CBF in absolute terms but is valid for determining relative changes in CBF during moderate flow increases (Fabricius and Lauritzen 1996). Stimulation-evoked CBF increases (ΔCBF = CBF response − CBF baseline) were normalized to the immediately preceding 20 s baseline (ΔCBF/CBF baseline). ΔCBF was measured as the area under the curve (AUC) and logarithmized for statistical purposes. After statistical analysis, CBF data were back-transformed, and ΔCBF is presented as AUC ± SEM.

### Local Tissue Oxygen Partial Pressure

The oxygen electrode was placed in the same lobe and at the same cortical depth as the electrode for electrophysiological recordings (∼0.1 mm apart). Local tpO2 was recorded with a modified Clark-type polarographic oxygen microelectrode (OX-10, Unisense A/S, Aarhus, Denmark). The small tip size (10 μm) assured reliable tpO2 measurements, and its built-in guard cathode removed all oxygen from the electrolyte reservoir. The field of sensitivity of the oxygen electrode is a sphere of 2× the tip diameter, that is, 4189 μm3. Before and after each experiment, each electrode was calibrated in air-saturated oxygen-free saline (0.9% at 37°C) with reproducible oxygen measurements. The oxygen electrodes were connected to a high-impedance pico-ampere meter (PA 2000; Unisense A/S), which sensed the currents of the oxygen electrodes. Signals were A/D-converted and recorded at 10 Hz (Power 1401 A/D converter running with Spike2 software; Cambridge Electronic Design). Current recordings were transformed to millimeter of mercury using the calibrations with saturated and oxygen-free standard solutions.

### Calculation of CMRO2

Considering that capillaries are parallel and non-communicating, it has been proposed that tissue oxygenation during functional stimulation may be altered through changes in the capillary flow without obvious changes in CBF (Jespersen and Ostergaard 2012). CMRO2 responses were calculated off-line from simultaneously obtained recordings of tpO2 and CBF, as described by Gjedde et al. (2005). The relationship between these 3 variables is given by

$tpO2=P502CaCBFCMRO2−1h−2CMRO23L,$
where P50 is the half-saturation tension of the oxygen–hemoglobin dissociation curve, h is the Hill coefficient of the same dissociation curve, Ca is the arterial oxygen concentration, and L is the effective diffusion coefficient of oxygen in brain tissue (Gjedde et al. 2005). The parameter L was determined using baseline CBF and CMRO2 values measured in rats anesthetized with α-chloralose—53 mL·100 g−1·min−1 and 219 μmol·100 g−1·min−1, respectively (Zhu et al. 2002)—and the average baseline oxygen tension for the animals included in this study. By this method, L was found to be 11.3 μmol·100 g−1·min−1·mmHg−1 for standard values of P50 = 36.0 mmHg; h = 2.7; and Ca = 8 μmol mL−1. As for CBF, the stimulation-evoked rises in CMRO2 (ΔCMRO2 = CMRO2 response − CMRO2 baseline) were normalized to the immediately preceding 20 s baseline, and ΔCMRO2 was measured as the AUC. ΔCMRO2 is presented as AUC ± SEM. All ΔCMRO2 data were log-transformed for statistical purposes and back-transformed after statistical analysis.

### Two-Photon Imaging

Ca2+ imaging was performed using a commercial two-photon microscope (SP5 multiphoton/confocal Laser Scanning Microscope; Leica, Germany), and a Mai Tai HP Ti:Sapphire laser (Millennia Pro, Spectra Physics, Sweden) with a ×20 1.0 NA-water-immersion objective (Leica). The excitation wavelength was set to 800 nm. The emitted light was filtered to retain both red and green light using a TRITC/FITC filter. The frame size was 128 × 128 pixels (85 ms frame−1).

The somatosensory cortex was surface-loaded with sulforhodamine 101 (SR101; Sigma–Aldrich) diluted to 1 mm in aCSF, which is specifically taken up by astrocytes (Nimmerjahn et al. 2004). Subsequently, the somatosensory cortex was covered with agarose and partly sealed with a glass cover slip. Following loading with SR101, the hemodynamic responses were measured as intrinsic optical signals (IOS) for localization of the active area for two-photon microscopy. We used a Leica microscope with ×4 magnification, allowing inclusion of the entire cranial window in the field of view. IOS images were obtained using a LED (light-emitting diode) light source with a green light filter and a fast-capturing camera (QuantEM 512SC) with a sampling rate of 28.8 Hz, before and during a 15-s 5-Hz whisker pad stimulation. The difference in absorption due to the shift in oxy- and deoxy-hemoglobin concentrations was calculated as described by Harrison et al. (2009).

The resultant IOS images were then used to guide loading of the membrane-permeant Ca2+ indicator Oregon Green Bapta-1/AM [Oregon Green BAPTA (OGB); Invitrogen, Molecular probes]. The indicator was dissolved in dimethyl sulfoxide plus 10% Pluronic F-127 (BASF Global) and diluted in aCSF to yield a final dye concentration of 0.8 mm. OGB was then pressure-injected (4–6 psi, 4 s; Pneumatic Pump; World Precision Instruments, Hertfordshire, United Kingdom) into the somatosensory cortex through a micropipette at a depth of 100–150 μm below the somatosensory cortex surface. The OGB stained all cellular structures in the targeted area.

### Image Analysis

We imaged the Ca2+ activity in astrocyte soma, neuron soma, and surrounding neuropil in layers 2/3 of the somatosensory cortex in response to whisker pad stimulation using the above-described electrophysiological protocol. Imaging parameters were adjusted to allow concomitant identification and imaging of several neuron and astrocyte somas in each frame. Whisker pad stimulation induced robust increases in fluorescence, reflecting Ca2+ activity in the tissue. The Ca2+ activity was compartmentalized based on the labeling pattern; astrocytes were labeled with both SR101 and OGB, giving rise to yellow cell structures, whereas neurons were labeled only with OGB and were thus green (Nimmerjahn et al. 2004).

Of the 10 mice imaged, 5 were treated with THIP and 5 with Zolpidem. Fluorescent Ca2+ transients were analyzed from 44 neuronal somas, 28 astrocyte somas, and 19 neuropil regions before and after applying increasing concentrations of THIP, and from 50 neuronal somas, 29 astrocyte somas, and 20 neuropil regions before and after applying increasing concentrations of Zolpidem. Thus, a total of 190 regions of interests (ROIs) were analyzed. The average fluorescence in the ROIs was calculated using Leica Application Suite Advanced Fluorescence (LAS-AF) v. 2.6.0.7266 (Leica Microsystems) and exported as a text file. For each ROI and each stimulation period, the average Ca2+ fluorescence during the 35-s period immediately preceding stimulation (F0) was used as baseline. At every time-point during stimulation, the evoked Ca2+ fluorescence signal (FS) was monitored and expressed as a percent of baseline (F0; ΔF = (FSF0)100/F0). The Ca2+ response was taken as the integral of the Ca2+ fluorescence signal above baseline, from the start of stimulation until the first time-point at which the signal decreased below baseline after the end of stimulation. The trapezoidal numerical algorithm (Matlab™) was used to calculate the integral of the Ca2+ response. In this manner, we determined the Ca2+ responses to every stimulation frequency in every ROI before and during drug treatment. Finally, we calculated the mean Ca2+ responses for each cellular compartment, that is, neurons, astrocytes, and neuropil for each drug concentration. All data concerning Ca2+ transients are presented as mean ± SEM.

For illustrative purposes, we have provided 2 examples of “difference images”, which show the Ca2+ activity in the entire field of view during a 2-Hz stimulation. To evaluate all responding cells in the field of view—including neuronal and astrocyte somas and the responding neuropil—the images obtained during baseline and during a 2-Hz stimulation were time-averaged pixel-by-pixel. The time-averaged baseline and 2-Hz stimulation images were aligned and subtracted, and the subtracted image was normalized to the time-averaged baseline image. Thus, the images show the Ca2+ activity during control conditions and following treatment with either THIP or Zolpidem.

### Drugs

In the present study, extrasynaptic inhibition was increased using THIP (Gaboxadol, 4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin-3-ol; kindly supplied by Bjarke Ebert of Lundbeck A/S, Copenhagen, Denmark), and stimulation-induced synaptic inhibition was increased using Zolpidem (Z-103; Sigma–Aldrich). Both substances were applied topically by placing droplets of the dissolved compound on the agarose covering the cortex. A total of 13 animals were treated with THIP and a total of 15 animals were treated with Zolpidem. THIP was dissolved in aCSF and applied at concentrations of 1, 10, and 100 to 13, 13, and 10 animals, respectively. Zolpidem was initially dissolved in DMSO and subsequently diluted with artificial cerebrospinal fluid (aCSF) giving a concentration of <0.1% DMSO in the final solution. Zolpidem was applied at concentrations of 0.5, 1, and 10 to 15, 15, and 13 animals, respectively.

### Statistical Analysis

The effects of THIP and Zolpidem were analyzed using mixed-effects analysis with R (Development Core Team 2010; R Foundation for Statistical Computing, Vienna, Austria). In all cases, we initially tested whether there was an overall effect of drug treatment on a specific parameter, that is, did the overall drug treatment differ from control values. If applicable we continued testing the individual drug concentrations separately compared with control values. Subsequently, paired two-tailed t-tests (Excel, Microsoft Office) were used to test for significance at specific stimulation frequencies. The significance level was set at α = 0.05. The results of the statistical analysis are summarized in Table 1.

Table 1

Summary of statistical analysis of the effect of drug treatment

Parameter Statistical model P-value N
Mixed-effects model

THIP
LFP (μm
1–100 df = 228
t-value = 5.303
<0.005 13
CBF (μm
1–100 df = 126
t-value = −3.466104
<0.005
1 df = 67
t-value = 4.795077
<0.0005
10 df = 66
t-value = 3.214227
<0.005
100 df = 49
t-value = −0.32275
<0.005
CMRO2m
1–100 df = 128
t-value = −3.64034
<0.005
1 df = 68
t-value = 3.23160
<0.005
10 df = 63
t-value = 3.07367
<0.005
100 df = 52
t-value = 4.25777
<0.0005
Ca2+ (overall drug effect) (μm
Neuronal somas
1–100
df = 363
t-value = −7.568
<0.05
Neuropile
1–100
df = 91
t-value = −3.161
<0.005
Astrocyte somas
1–100
df = 1214
t-value = −3.363
<0.0005
Zolpidem
LFP (μm
0.5–10 df = 251
t-value = 2.8496
<0.005 15
CBF (μm
0.5–10 df = 159
t-value = 3.667086
<0.0005 10
0.5 df = 85
t-value = 0.988284
= 0.3258 10
1 df = 74
t-value = 4.120925
<0.0005 10
10 df = 71
t-value = 2.496632
<0.05
CMRO2m
0.5–10 df = 161
t-value = 2.50102
<0.05 10
0.5 df = 86
t-value = 0.72891
= 0.4680 10
1 df = 73
t-value = 2.25672
<0.05 10
10 df = 73
t-value = −2.21085
<0.05
Ca2+ (overall drug effect) (μm
Neuronal somas
0.5–10
df = 82
t-value = −2.361
<0.05
Neuropile
0.5–10
df = 82
t-value = −4.052
<0.0005
Astrocyte
Somas
0.5–10
df = 82
t-value = −4.182
<0.0005
Ca2+ (0.5 μm
Neuronal somas df = 42
t-value = 2.805
<0.05
Neuropile df = 42
t-value = 0.8778
0.385
Astrocyte somas df = 42
t-value = 0.142
0.888
Parameter Statistical model P-value N
Mixed-effects model

THIP
LFP (μm
1–100 df = 228
t-value = 5.303
<0.005 13
CBF (μm
1–100 df = 126
t-value = −3.466104
<0.005
1 df = 67
t-value = 4.795077
<0.0005
10 df = 66
t-value = 3.214227
<0.005
100 df = 49
t-value = −0.32275
<0.005
CMRO2m
1–100 df = 128
t-value = −3.64034
<0.005
1 df = 68
t-value = 3.23160
<0.005
10 df = 63
t-value = 3.07367
<0.005
100 df = 52
t-value = 4.25777
<0.0005
Ca2+ (overall drug effect) (μm
Neuronal somas
1–100
df = 363
t-value = −7.568
<0.05
Neuropile
1–100
df = 91
t-value = −3.161
<0.005
Astrocyte somas
1–100
df = 1214
t-value = −3.363
<0.0005
Zolpidem
LFP (μm
0.5–10 df = 251
t-value = 2.8496
<0.005 15
CBF (μm
0.5–10 df = 159
t-value = 3.667086
<0.0005 10
0.5 df = 85
t-value = 0.988284
= 0.3258 10
1 df = 74
t-value = 4.120925
<0.0005 10
10 df = 71
t-value = 2.496632
<0.05
CMRO2m
0.5–10 df = 161
t-value = 2.50102
<0.05 10
0.5 df = 86
t-value = 0.72891
= 0.4680 10
1 df = 73
t-value = 2.25672
<0.05 10
10 df = 73
t-value = −2.21085
<0.05
Ca2+ (overall drug effect) (μm
Neuronal somas
0.5–10
df = 82
t-value = −2.361
<0.05
Neuropile
0.5–10
df = 82
t-value = −4.052
<0.0005
Astrocyte
Somas
0.5–10
df = 82
t-value = −4.182
<0.0005
Ca2+ (0.5 μm
Neuronal somas df = 42
t-value = 2.805
<0.05
Neuropile df = 42
t-value = 0.8778
0.385
Astrocyte somas df = 42
t-value = 0.142
0.888

Note: The results of mixed-effects model statistical analysis for the effects of THIP or Zolpidem on all investigated parameters, including LFP, ΔCBF, ΔCMRO2, and Ca2+ responses. We investigated whether there was an overall effect of treatment with either THIP (1–100 μm) or Zolpidem (0.5–10 μm) as compared with control values. If applicable, we next tested for effect of the individual concentrations if the overall drug effect was significantly different from the control values.

Increases in CBF (ΔCBF) and CMRO2 (ΔCMRO2) were normalized to the immediately preceding 20 s baseline. Ca2+ response transients were normalized to the immediately preceding 35 s baseline. LFP was measured in all experiments, and the summed field potential (ΣLFP) was correlated to ΔCBF, ΔCMRO2, and Ca2+ transients. ΔCBF and ΔCMRO2 data were logarithmized to ensure a normal distribution of the residuals. Data are presented as mean value ± SEM.

## Results

Fig. 1A shows the basic experimental setup with all the hardware components used for data acquisition, and Fig. 1BG shows the stimulation-evoked rises in CBF, CMRO2, and LFPs. Using two-photon microscopy and a microelectrode, we recorded fluorescent Ca2+ transients in neuropil, and neuronal and astrocyte soma, and LFPs in layers 2/3 of mouse somatosensory cortex (Fig. 1HN). Whisker pad stimulation resulted in frequency-dependent variations in LFP, with ΣLFP reaching the maximum at a 3-Hz stimulation frequency and decreasing at higher frequencies (Fig. 1D,G). Similar patterns were observed for the stimulation-evoked rises in CBF and CMRO2, both of which plateaued at 2–5 Hz (Fig. 1E,F). The stimulation-induced hemodynamic response was mapped using IOS. Fig. 1H shows the craniotomy and exposed brain area during control conditions, and functions as an anatomical image. IOS shows darkening of the parenchyma in the activated area when blood flow increases during whisker pad stimulation (Fig. 1I). Fluorophores for two-photon microscopy were injected into the activated (darkened) region. Fig. 1J shows 2 simultaneously obtained representative two-photon images. Astrocyte somas and processes are labeled with the astrocyte-specific marker SR101 (Fig. 1J, left panel) (Nimmerjahn et al. 2004), whereas neuronal soma, astrocyte soma, and neuropil are all labeled with OGB (middle panel) (Stosiek et al. 2003). The merged image to the right shows layer 2/3 neurons labeled with OGB and astrocytes labeled with both OGB and SR101, as indicated by the arrowheads. Neuropil was defined as areas lacking morphological characteristics of either neuronal cell somas or astrocyte cell structures. Based on the morphology and the double staining of the cellular elements, ROIs were allocated to neuropil, neuronal somatic, or astrocyte somatic compartments (Fig. 1KM). Fig. 1N shows typical frequency-dependent cytosolic Ca2+ transients in neuropil, and neuronal and astrocyte somas during whisker pad stimulation. Ca2+ transients were significantly larger in the neuropil than in the neuronal or astrocyte somas and were greatly dependent on stimulation frequency, with peak responses at 2 Hz and smaller responses at lower or higher stimulation frequencies.

Figure 1.

Figure 1.

### Extrasynaptic and Synaptic GABAAR Activities Differently Affect Synaptic Activity

Whisker pad stimulation results in activity in somatosensory afferents, which project via the thalamus to layer IV of the somatosensory cortex and, after a few milliseconds, to the layer 2/3 pyramidal cells (Armstrong-James et al. 1992). The synaptic input produces a large EPSP due to the activation of pyramidal neurons in all cortical layers, which was recorded as a negative extracellular LFP. LFP amplitude was greatly modified by both THIP and Zolpidem (Fig. 2A,B). At 1 μm, THIP had a clear bidirectional effect on LFP amplitude, increasing it at low stimulation frequencies and decreasing it at higher frequencies. THIP concentrations of ≥10 μm decreased LFP amplitude at all stimulation frequencies (Fig. 2A). Similarly to THIP, Zolpidem concentrations of 0.5 and 1 μm increased LFP amplitude in response to low stimulation frequencies, without significantly reducing LFP amplitude at higher stimulation frequencies (Fig. 2B). Increasing the Zolpidem concentration to 10 μm had no effect on LFP amplitude at low stimulation frequencies but significantly decreased LFP at higher stimulation frequencies. We conclude that both THIP and Zolpidem enhanced postsynaptic excitation at low concentrations and low stimulation frequencies, whereas suppressing it at higher drug concentrations and/or at intermediate or high stimulation frequencies. These results suggest that small increases in the activity of synaptic and extrasynaptic GABAARs may amplify transient low-frequency somatosensory input and filter out input of intermediate and higher frequencies.

Figure 2.

Bidirectional modulation of synaptic activity, and neurovascular and metabolic coupling. (A-B) Summarized and averaged stimulation-evoked LFP amplitudes at stimulation frequencies of 0.5, 1.0, 2.0, 3.0, and 5.0 Hz under control conditions (gray line), or following treatment with THIP (A) at 1 μm (yellow line), 10 μm (light green line), and 100 μm (dark green line) or treatment with Zolpidem (B) at 0.5 μm (yellow line), 1 μm (orange line), and 10 μm (red line). Evoked LFP responses are presented as the size of the negative amplitude (mV) for each stimulation frequency. THIP at 1 μm increased LFP amplitude at 0.5 and 1.0 Hz but decreased it at higher frequencies (2.0–5.0 Hz). Zolpidem 0.5 μm increased LFP amplitude at 0.5 and 1.0 Hz but decreased LFP at higher stimulation frequencies and drug concentrations. Data are shown as mean amplitude (mV) ± SEM. Color-coordinated asterisks indicate significance levels of the corresponding drug concentrations. *P < 0.05 and **P < 0.005 in AF. (C and D) Summarized and averaged stimulation-evoked ΔCBF responses under control conditions and following application of THIP or Zolpidem as in A and B. Evoked ΔCBF responses are normalized to the immediately preceding baseline and presented as the AUC ± SEM for each stimulation frequency. Activation of extrasynaptic GABAARs by low concentrations of THIP (C) increased the stimulation-evoked ΔCBF at low stimulation frequencies (<2 Hz). Higher THIP concentrations decreased the ΔCBF responses at higher stimulation frequencies (≥2 Hz). Zolpidem (D) at low concentrations modulated the evoked ΔCBF responses, inducing slight response increases at low stimulation frequencies, but having no effect at higher stimulation frequencies. Increasing Zolpidem concentration decreased responses at high stimulation frequencies. (E and F) Summarized and averaged stimulation-evoked ΔCMRO2 responses during control conditions and following application of THIP or Zolpidem as in A and B. Evoked ΔCMRO2 are calculated as for ΔCBF in C and D. At low concentrations, THIP (E) bidirectionally modulated ΔCMRO2. Increasing THIP concentration reduced responses at higher stimulation frequencies (≥2 Hz). Zolpidem substantially modulated ΔCMRO2. At intermediate concentration, Zolpidem increased the ΔCMRO2 response at low stimulation frequency.

Figure 2.

Bidirectional modulation of synaptic activity, and neurovascular and metabolic coupling. (A-B) Summarized and averaged stimulation-evoked LFP amplitudes at stimulation frequencies of 0.5, 1.0, 2.0, 3.0, and 5.0 Hz under control conditions (gray line), or following treatment with THIP (A) at 1 μm (yellow line), 10 μm (light green line), and 100 μm (dark green line) or treatment with Zolpidem (B) at 0.5 μm (yellow line), 1 μm (orange line), and 10 μm (red line). Evoked LFP responses are presented as the size of the negative amplitude (mV) for each stimulation frequency. THIP at 1 μm increased LFP amplitude at 0.5 and 1.0 Hz but decreased it at higher frequencies (2.0–5.0 Hz). Zolpidem 0.5 μm increased LFP amplitude at 0.5 and 1.0 Hz but decreased LFP at higher stimulation frequencies and drug concentrations. Data are shown as mean amplitude (mV) ± SEM. Color-coordinated asterisks indicate significance levels of the corresponding drug concentrations. *P < 0.05 and **P < 0.005 in AF. (C and D) Summarized and averaged stimulation-evoked ΔCBF responses under control conditions and following application of THIP or Zolpidem as in A and B. Evoked ΔCBF responses are normalized to the immediately preceding baseline and presented as the AUC ± SEM for each stimulation frequency. Activation of extrasynaptic GABAARs by low concentrations of THIP (C) increased the stimulation-evoked ΔCBF at low stimulation frequencies (<2 Hz). Higher THIP concentrations decreased the ΔCBF responses at higher stimulation frequencies (≥2 Hz). Zolpidem (D) at low concentrations modulated the evoked ΔCBF responses, inducing slight response increases at low stimulation frequencies, but having no effect at higher stimulation frequencies. Increasing Zolpidem concentration decreased responses at high stimulation frequencies. (E and F) Summarized and averaged stimulation-evoked ΔCMRO2 responses during control conditions and following application of THIP or Zolpidem as in A and B. Evoked ΔCMRO2 are calculated as for ΔCBF in C and D. At low concentrations, THIP (E) bidirectionally modulated ΔCMRO2. Increasing THIP concentration reduced responses at higher stimulation frequencies (≥2 Hz). Zolpidem substantially modulated ΔCMRO2. At intermediate concentration, Zolpidem increased the ΔCMRO2 response at low stimulation frequency.

### Effects of Synaptic and Extrasynaptic GABAAR Activity on Stimulus-Evoked CBF and CMRO2 Responses

We next explored the influence of GABAAR activity on rises in blood flow and CMRO2, that is, the mechanisms that underlie functional neuroimaging signals. First, we examined the effect of THIP, which modulates tonic inhibition through extrasynaptic GABAAR (Fig. 2C,E). Overall, THIP induced significant effects on both evoked ΔCBF and ΔCMRO2 compared with that in control conditions (Fig. 2C,E). Specifically, 1 μm THIP significantly increased the ΔCBF at low stimulation frequencies but had no effect at intermediate and high stimulation frequencies; THIP at 10 and 100 μm significantly decreased ΔCBF at high stimulation frequencies; and THIP at 10 μm showed a tendency towards increasing ΔCBF at low stimulation frequencies but the effect was not significant (Fig. 2C). Thus, the initial increase in synaptic activity (LFP) at low stimulation frequencies induced by 1 μm THIP is reflected as an increased ΔCBF. The depressed synaptic activity (LFP) seen at intermediate or high stimulation frequencies and at increased THIP concentrations was reflected as reduced ΔCBF. THIP also significantly modulated ΔCMRO2. At 1 μm, THIP increased ΔCMRO2 at low stimulation frequencies and reduced ΔCMRO2 at high stimulation frequencies. Increasing the THIP concentration (10–100 μm) tended to increase ΔCMRO2 at low stimulation frequencies, but the effect was not significant, whereas high THIP concentrations significantly decreased ΔCMRO2 at higher stimulation frequencies.

In contrast to THIP, Zolpidem has no direct effects. Zolpidem modulates GABAAR activity by increasing the gain of GABA released at the inhibitory synapse during stimulation (Krogsgaard-Larsen and Johnston 1978). We found that Zolpidem also had an overall significant effect on both ΔCBF and ΔCMRO2 (Fig. 2D,F). At 1 μm, Zolpidem increased ΔCBF at 1 Hz, but had no significant effects on ΔCBF at other stimulation frequencies, whereas at 0.5 μm Zolpidem had no significant effect on ΔCBF at any frequency (Fig. 2D). Increasing the Zolpidem concentration to 10 μm reduced ΔCBF responses to high stimulation frequencies, but did not significantly affect ΔCBF at low stimulation frequencies. Overall, Zolpidem also significantly modulated the ΔCMRO2 response. Similar to its effect on ΔCBF, Zolpidem significantly increased ΔCMRO2 at 1 μM and 1 Hz, whereas 0.5 and 10 μm Zolpidem had no significant effect on ΔCMRO2 at any frequency. Thus, at low stimulation frequencies and low concentrations, Zolpidem had similar effects on ΔCMRO2, ΔCBF, and synaptic activity.

In summary, THIP and to a lesser extent Zolpidem bidirectionally modulated synaptic activity, ΔCBF, and ΔCMRO2. Table 2 presents the power analyses of electrical synaptic activity, CBF, and CMRO2. We demonstrate that both THIP and Zolpidem significantly decrease baseline CBF, but not baseline CMRO2. Zolpidem does not significantly affect the power of synaptic activity in baseline conditions, whereas THIP at 10 and 100 μm significantly decreases it. At low drug concentrations and low stimulation frequencies, these responses were augmented. At high drug concentrations and high stimulation frequencies, both ΔCBF and LFP amplitude were significantly decreased. THIP increased synaptic activity at low concentrations and low frequency, while inhibiting LFP at both intermediate concentrations (high frequencies) and high concentrations (all frequencies). At high concentrations, THIP reduced both ΔCMRO2 and ΔCBF, but only at high stimulation frequencies. Zolpidem increased synaptic activity, ΔCBF, and ΔCMRO2 at low stimulation frequencies and at low concentrations, whereas it significantly decreased synaptic activity and ΔCBF, but not ΔCMRO2, at high concentrations and higher stimulation frequencies. Thus, the drug effects we observed were dependent on both concentration and frequency. This suggests that small increases in synaptic and extrasynaptic GABAAR activity may amplify vascular and metabolic responses to transient low-frequency somatosensory inputs, while reducing vascular and metabolic signals evoked by intermediate and high-frequency excitatory inputs.

Table 2

Summary of the effects on baseline values of synaptic activity, CBF, and CMRO2

THIP (μm1–100 (overall drug effect) 10 100

Synaptic activity P < 0.001 P = 0.298 P < 0.001 P < 0.01
CBF
Percent change
P < 0.0001 P < 0.005
20 ± 6.6%
P < 0.0005
26 ± 6.4%
P < 0.0001
23 ± 8.3%
CMRO2 P = 0.0803 — — —
Zolpidem (μm0.5–10 (overall drug effect) 0.5 10

Synaptic activity P = 0.1289 — — —
CBF
Percent change
P < 0.0001 P = 0.071
12 ± 6.2%
P < 0.001
25 ± 4.1%
P < 0.001
31 ± 5.6%
CMRO2 P = 0.1378 — — —
THIP (μm1–100 (overall drug effect) 10 100

Synaptic activity P < 0.001 P = 0.298 P < 0.001 P < 0.01
CBF
Percent change
P < 0.0001 P < 0.005
20 ± 6.6%
P < 0.0005
26 ± 6.4%
P < 0.0001
23 ± 8.3%
CMRO2 P = 0.0803 — — —
Zolpidem (μm0.5–10 (overall drug effect) 0.5 10

Synaptic activity P = 0.1289 — — —
CBF
Percent change
P < 0.0001 P = 0.071
12 ± 6.2%
P < 0.001
25 ± 4.1%
P < 0.001
31 ± 5.6%
CMRO2 P = 0.1378 — — —

Note: Results of mixed-effects model statistical analysis for the effects of THIP or Zolpidem on the baseline conditions of CBF, CMRO2, and synaptic activity. We investigated whether there was an overall effect of treatment with either THIP (1–100 μm) or Zolpidem (0.5–10 μm) as compared with control values. If applicable, we next tested for effect of the individual concentrations if the overall drug effect was significantly different from the control values.

### Coupling of Hemodynamic and Metabolic Responses to Total Synaptic Activity

To examine whether gain modulation of the excitatory synaptic input could explain the observed effects of increased GABAAR activity on ΔCBF or ΔCMRO2, we plotted synaptic activity (ΣLFP) versus ΔCBF or ΔCMRO2 across all frequencies (Tables 3 and 4) for each drug concentration. Under control conditions, the relation between ΔCBF and ΣLFP was practically linear, confirming the dependency of ΔCBF on postsynaptic activity (Fig. 3A,B). Likewise, under control conditions, the relationship between ΔCMRO2 and ΣLFP was linear for THIP and Zolpidem (Fig. 3C,D and Table 4). The correlations (Spearman's correlation coefficient; ρ) of ΣLFP with ΔCBF and ΔCMRO2 decreased with increasing concentrations of both THIP and Zolpidem. This reduction in ρ could be due to the reduction in the range of ΣLFP values available for correlation analysis at the highest drug concentration, which was especially evident for THIP. However, correlation analyses excluding the highest concentrations of both drugs showed the same pattern of decrease with increasing drug concentrations, indicating gradual uncoupling of the responses as GABAAR activity increased (Tables 3 and 4).

Table 3

Correlation between ΔCBF and ΣLFP in the presence of THIP or Zolpidem

ρ SDρ Slope SDslope
THIP (μm
0.925 0.046 2.430 1.056
0.700* 0.261 1.877 1.100
10 0.263** 0.430 2.025 1.699
100 0.280** 0.286 0.408 1.019
Zolpidem (μm
0.725 0.116 2.214 1.030
0.5 0.560 0.462 1.697 2.238
0.363* 0.362 1.835 2.812
10 −0.013** 0.494 0.168 0.902
ρ SDρ Slope SDslope
THIP (μm
0.925 0.046 2.430 1.056
0.700* 0.261 1.877 1.100
10 0.263** 0.430 2.025 1.699
100 0.280** 0.286 0.408 1.019
Zolpidem (μm
0.725 0.116 2.214 1.030
0.5 0.560 0.462 1.697 2.238
0.363* 0.362 1.835 2.812
10 −0.013** 0.494 0.168 0.902

Note: Correlation coefficients (ρ) and the standard deviation (SDρ) were calculated over all stimulation frequencies for each drug, concentration, and mouse and are presented as the averages for each concentration. The correlation coefficients (ρ) were then analyzed for drug effect. *P < 0.05; **P < 0.005. The slope and standard deviation (SDslope) of each regression line were calculated and analyzed for drug effects. Increasing concentrations of both THIP and Zolpidem reduced the correlation between ΔCBF and ΣLFP, indicating that increased GABAAR activity weakened the tight correlation between these 2 parameters. This is reflected in the decrease of the slope obtained by linear regression of ΔCBF onto ΣLFP. Note the strong linearity between ΔCBF and ΣLFP during control conditions and the similar decreases in correlation strength for both THIP and Zolpidem.

Table 4

Correlation between ΔCMRO2 and ΣLFP in the presence of THIP and Zolpidem

ρ SDρ Slope SDslope
THIP (μm
0 0.825 0.158 1.442 0.556
1 0.413 0.666 0.650 0.347
10 0.025*** 0.504 0.886 2.577
100 −0.02* 0.491 −2.026 3.894
Zolpidem (μm
0 0.800 0.141 1.069 0.857
0.5 0.450 0.517 1.381 1.827
1 0.288* 0.503 1.114 1.100
10 0.588 0.253 1.045 0.888
ρ SDρ Slope SDslope
THIP (μm
0 0.825 0.158 1.442 0.556
1 0.413 0.666 0.650 0.347
10 0.025*** 0.504 0.886 2.577
100 −0.02* 0.491 −2.026 3.894
Zolpidem (μm
0 0.800 0.141 1.069 0.857
0.5 0.450 0.517 1.381 1.827
1 0.288* 0.503 1.114 1.100
10 0.588 0.253 1.045 0.888

Note: Correlation coefficients (ρ) and the standard deviation (SDρ) were calculated over all stimulation frequencies for each drug, concentration, and mouse and are presented as the averages for each concentration. The correlation coefficients (ρ) were then analyzed for drug effect. *P < 0.05; **P < 0.005; ***P < 0.0005. The slope and standard deviation (SDslope) of the regression lines were calculated and analyzed for drug effects. Increasing concentrations of both THIP and Zolpidem reduced the correlation between ΔCMRO2 and ΣLFP, as indicated by the slope reduction obtained by linear regression of ΔCMRO2 onto ΣLFP. Linearity was found between ΔCMRO2 and ΣLFP during control conditions, but the correlation was less sensitive to drug treatment compared with the correlation strength for ΔCBF and ΣLFP.

Figure 3.

Neurovascular and metabolic coupling is deregulated during the activation of both extrasynaptic and synaptic GABAAR. (A and B) Summarized evoked responses of ΔCBF before and after application of THIP or Zolpidem as a function of the total synaptic activity ΣLFP (calculated as LFPamplitude × stimulation frequency). (C and D) Summarized evoked responses of ΔCMRO2 before and after application of either THIP or Zolpidem as a function of ΣLFP. ΔCBF and ΔCMRO2 responses are presented as AUC ± SEM and normalized to the immediately preceding baseline. The graphs show linear correlations of ΣLFP with both ΔCBF and ΔCMRO2 responses under control conditions. The correlation analysis results are summarized in Table 3.

Figure 3.

Neurovascular and metabolic coupling is deregulated during the activation of both extrasynaptic and synaptic GABAAR. (A and B) Summarized evoked responses of ΔCBF before and after application of THIP or Zolpidem as a function of the total synaptic activity ΣLFP (calculated as LFPamplitude × stimulation frequency). (C and D) Summarized evoked responses of ΔCMRO2 before and after application of either THIP or Zolpidem as a function of ΣLFP. ΔCBF and ΔCMRO2 responses are presented as AUC ± SEM and normalized to the immediately preceding baseline. The graphs show linear correlations of ΣLFP with both ΔCBF and ΔCMRO2 responses under control conditions. The correlation analysis results are summarized in Table 3.

For THIP, we observed an overall significant effect of increasing drug concentration on ρ for both &triangle;CMRO2 and &triangle;CBF versus ΣLFP (Fig. 3A,C and Tables 3 and 4). The correlation between ΔCBF and ΣLFP showed a clear concentration-dependent effect of THIP (Table 3), whereas the correlation between ΔCMRO2 and ΣLFP was only affected significantly by THIP concentrations of ≥10 μm (Table 4). These findings suggest that uncoupling between synaptic activity (ΣLFP), ΔCBF, and ΔCMRO2 may occur with only slight variations in the level of tonic inhibition, that is, in the level of activity at extrasynaptic GABAARs. Zolpidem also induced overall significant effects on the correlations between ΣLFP and ΔCBF (Fig. 3B and Table 3) and between ΣLFP and ΔCMRO2 (Table 4). These correlations were observed with 0.5 μm Zolpidem and decreased with 1 and 10 μm Zolpidem. These data suggest that increased activity at synaptic GABAARs may uncouple synaptic activity from ΔCMRO2 and ΔCBF but that the correlation between ΔCMRO2 and ΣLFP may be less sensitive to phasic than to tonic increases, that is, to extrasynaptic inhibition. This indicates that rises in synaptic GABAARs activity may modulate vascular and metabolic responses to somatosensory inputs in part by influencing excitatory inputs but that the effect is not fully explained by increased modulation of postsynaptic activity.

### Effects of GABAAR Activation on Stimulus-Evoked Cytosolic Ca2+ Signals

We next tested the hypothesis that Ca2+-dependent mechanisms explain the drug effects on ΔCBF, which are known to depend strongly on Ca2+-related mechanisms (Lauritzen 2005; Iadecola and Nedergaard 2007; Koehler et al. 2009). Fluorescent Ca2+ signals in astrocyte and neuronal soma, and neuropil ROIs varied depending on stimulation frequency, with peak responses at 2 Hz and smaller responses at lower and higher frequencies (Fig. 1N). Figures 4A and B show processed images of representative fluorescent Ca2+ signals from the mouse somatosensory cortex. The results of THIP treatment are summarized in Figure 4CE. THIP decreased the overall Ca2+ signal in neuronal soma (Fig. 4C), in neuropil (Fig. 4D), and in astrocyte soma (Fig. 4E) in a concentration-dependent manner. At 0.5 Hz, 1 μm THIP showed a tendency towards increasing the Ca2+ activity in the different ROI types, but this effect was not statistically significant.

Figure 4.

Neuroimaging of evoked Ca2+ activity during the activation of extrasynaptic GABAAR by THIP. (A) Raw images obtained during an experiment. The image to the far left shows labeling of astrocyte somas and cellular structures with the astrocyte-specific marker SR101. The middle image shows the same frame, with neuronal and astrocyte somas and neuropil labeled with OGB. The image to the right is a merged image of SR101 and OGB fluorescence, showing different labeling of astrocyte somas (yellow) and neuronal somas and neuropil (green). Scale bar = 25 μm. (B) Differential images obtained during 2-Hz stimulation, before and after application of 1–100 μm THIP, showing raw OGB fluorescence changes in the scanned area. Increasing THIP concentration decreased the average Ca2+ fluorescence in a concentration-dependent manner in neuronal soma, astrocyte soma, and neuropil, as illustrated by the rainbow scale. Scale bar = 25 μm. (C–E) Summarized and averaged Ca2+ activity in neuronal soma, astrocyte soma, and neuropil ROIs under control conditions (gray line), and following treatment with THIP at 1 (yellow line), 10 (light green line), and 100 μm (dark green line). Ca2+ activity is calculated as ΔF/F0 and presented as AUC ± SEM. THIP caused significant concentration- and frequency-dependent decreases in Ca2+ activity in neuronal soma (C), neuropil (D), and astrocyte soma (E). Color-coordinated asterisks indicate significance levels of the corresponding THIP concentrations. *P < 0.05, **P < 0.005, and ***P < 0.0005.

Figure 4.

Neuroimaging of evoked Ca2+ activity during the activation of extrasynaptic GABAAR by THIP. (A) Raw images obtained during an experiment. The image to the far left shows labeling of astrocyte somas and cellular structures with the astrocyte-specific marker SR101. The middle image shows the same frame, with neuronal and astrocyte somas and neuropil labeled with OGB. The image to the right is a merged image of SR101 and OGB fluorescence, showing different labeling of astrocyte somas (yellow) and neuronal somas and neuropil (green). Scale bar = 25 μm. (B) Differential images obtained during 2-Hz stimulation, before and after application of 1–100 μm THIP, showing raw OGB fluorescence changes in the scanned area. Increasing THIP concentration decreased the average Ca2+ fluorescence in a concentration-dependent manner in neuronal soma, astrocyte soma, and neuropil, as illustrated by the rainbow scale. Scale bar = 25 μm. (C–E) Summarized and averaged Ca2+ activity in neuronal soma, astrocyte soma, and neuropil ROIs under control conditions (gray line), and following treatment with THIP at 1 (yellow line), 10 (light green line), and 100 μm (dark green line). Ca2+ activity is calculated as ΔF/F0 and presented as AUC ± SEM. THIP caused significant concentration- and frequency-dependent decreases in Ca2+ activity in neuronal soma (C), neuropil (D), and astrocyte soma (E). Color-coordinated asterisks indicate significance levels of the corresponding THIP concentrations. *P < 0.05, **P < 0.005, and ***P < 0.0005.

The corresponding images and summarized results from mice treated with Zolpidem are shown in Figure 5AE. Zolpidem induced an overall significant effect on Ca2+ responses in neuronal soma (Fig. 5C), neuropil (Fig. 5D), and astrocyte soma (Fig. 5E). At 0.5 μm, Zolpidem augmented the Ca2+ signal across all frequencies in neuronal soma, without affecting the Ca2+ signal in either neuropil or astrocyte soma. Neither Zolpidem nor THIP affected the number of ROIs responding to stimulation following drug application (data not shown). Our data indicate that enhancement of phasic inhibition using low concentrations of Zolpidem significantly increased the Ca2+ response in neuronal soma, whereas high concentrations of Zolpidem and all concentrations of THIP reduced Ca2+ response in all cell structures and neuropil.

Figure 5.

Neuroimaging of evoked Ca2+ activity during the activation of synaptic GABAAR by Zolpidem. (A) Raw images obtained during an experiment. The image to the far left shows labeling of astrocyte somas and astrocyte cellular structures with the astrocyte-specific marker SR101. The middle image shows the same area, with neuronal and astrocyte somas and neuropil labeled with OGB. The image to the right is a merged image of SR101 and OGB fluorescence, showing different labeling of astrocyte somas (yellow) and neuronal somas and neuropil (green). (B) Differential images of OGB fluorescence obtained during 2-Hz stimulation, before and after application of 0.5, 1, and 10 μm Zolpidem, showing responses of the whole scanned area. The changes in average Ca2+ fluorescence are illustrated by the rainbow scale. Scale bar = 25 μm. (CE) Summarized and averaged Ca2+ activity in neuronal soma, astrocyte soma, and neuropil ROIs under control conditions (gray line), and following treatment with Zolpidem at 0.5 μm (yellow line), 1 μm (orange line), and 10 μm (red line). Ca2+ activity is calculated as ΔF/F0 and presented as the AUC ± SEM. While 0.5 μm Zolpidem increased neuronal Ca2+ responses significantly at all stimulation frequencies, the higher drug concentrations reduced Ca2+ responses in neuronal somas (C). In neuropil (D) and astrocyte somas (E), Zolpidem reduced Ca2+ responses in a concentration- and frequency-dependent manner. Color-coordinated asterisks indicate significance levels of the corresponding Zolpidem concentrations. *P < 0.05, **P < 0.005, and ***P < 0.0005.

Figure 5.

Neuroimaging of evoked Ca2+ activity during the activation of synaptic GABAAR by Zolpidem. (A) Raw images obtained during an experiment. The image to the far left shows labeling of astrocyte somas and astrocyte cellular structures with the astrocyte-specific marker SR101. The middle image shows the same area, with neuronal and astrocyte somas and neuropil labeled with OGB. The image to the right is a merged image of SR101 and OGB fluorescence, showing different labeling of astrocyte somas (yellow) and neuronal somas and neuropil (green). (B) Differential images of OGB fluorescence obtained during 2-Hz stimulation, before and after application of 0.5, 1, and 10 μm Zolpidem, showing responses of the whole scanned area. The changes in average Ca2+ fluorescence are illustrated by the rainbow scale. Scale bar = 25 μm. (CE) Summarized and averaged Ca2+ activity in neuronal soma, astrocyte soma, and neuropil ROIs under control conditions (gray line), and following treatment with Zolpidem at 0.5 μm (yellow line), 1 μm (orange line), and 10 μm (red line). Ca2+ activity is calculated as ΔF/F0 and presented as the AUC ± SEM. While 0.5 μm Zolpidem increased neuronal Ca2+ responses significantly at all stimulation frequencies, the higher drug concentrations reduced Ca2+ responses in neuronal somas (C). In neuropil (D) and astrocyte somas (E), Zolpidem reduced Ca2+ responses in a concentration- and frequency-dependent manner. Color-coordinated asterisks indicate significance levels of the corresponding Zolpidem concentrations. *P < 0.05, **P < 0.005, and ***P < 0.0005.

### Coupling Between Synaptic Activity and the Total Ca2+ Response

To further elucidate the different effects between synaptic or extrasynaptic GABAAR activity on Ca2+ responses, we investigated the correlation between the Ca2+ responses in neuronal soma, neuropil, and astrocyte soma and the total synaptic activity ΣLFP. Extracellular LFP recordings were made simultaneously with two-photon Ca2+ imaging. Under control conditions, ΣLFP correlated strongly with Ca2+ responses (Spearman's correlation coefficient, ρ) in the neuropil, neuronal soma, and astrocyte soma (Fig. 6AF and Table 5). For all 3 cellular compartments, the correlation between Ca2+ responses and ΣLFPs decreased significantly in the presence of THIP (Table 5). In neuropil, 100 μm THIP significantly decreased the correlation between Ca2+ response and ΣLFP; in neuronal soma, the correlation was affected by 1 and 10 μm THIP; and at 10 μm, THIP significantly decreased the correlation between Ca2+ response in astrocyte soma and ΣLFP (Table 5). These data indicate that increased tonic inhibition can decouple Ca2+ responses in neuropil and neuronal and astrocyte somas from synaptic activity.

Table 5

Correlation between Ca2+ responses and ΣLFP in the presence of THIP and Zolpidem

Neuropil

Neuronal somas

Astrocyte somas

ρ SDρ Slope SDslope ρ SDρ Slope SDslope ρ SDρ Slope SDslope
THIP (μm
0 0.750 0.155 1.338 0.122 0.700 0.200 1.291 0.062 0.725 0.150 1.252 0.265
1 0.080 0.593 0.785 0.264 0.020* 0.432 0.776 0.455 0.260 0.513 0.923 0.081
10 −0.220** 0.782 0.182 0.395 −0.280* 0.449 −0.330 0.839 0.060* 0.615 0.916 0.408
100 0.04** 0.261 0.455 0.331 −0.360 0.462 −0.540 0.593 0.480 0.311 0.752 0.328
Zolpidem (μm
0 0.725 0.171 1.729 0.219 0.700 0.200 2.210 0.338 0.725 0.171 1.305 0.346
0.5 0.420 0.383 1.701 0.153 0.340 0.351 3.787 0.175 0.06* 0.329 0.708 0.420
1 0.260 0.562 1.373 0.125 0.020 0.643 1.836 0.499 −0.180 0.402 −0.458 0.045
10 0.467 0.924 0.596 0.433 0.200 0.917 0.205 0.801 0.167 0.709 −0.694 0.448
Neuropil

Neuronal somas

Astrocyte somas

ρ SDρ Slope SDslope ρ SDρ Slope SDslope ρ SDρ Slope SDslope
THIP (μm
0 0.750 0.155 1.338 0.122 0.700 0.200 1.291 0.062 0.725 0.150 1.252 0.265
1 0.080 0.593 0.785 0.264 0.020* 0.432 0.776 0.455 0.260 0.513 0.923 0.081
10 −0.220** 0.782 0.182 0.395 −0.280* 0.449 −0.330 0.839 0.060* 0.615 0.916 0.408
100 0.04** 0.261 0.455 0.331 −0.360 0.462 −0.540 0.593 0.480 0.311 0.752 0.328
Zolpidem (μm
0 0.725 0.171 1.729 0.219 0.700 0.200 2.210 0.338 0.725 0.171 1.305 0.346
0.5 0.420 0.383 1.701 0.153 0.340 0.351 3.787 0.175 0.06* 0.329 0.708 0.420
1 0.260 0.562 1.373 0.125 0.020 0.643 1.836 0.499 −0.180 0.402 −0.458 0.045
10 0.467 0.924 0.596 0.433 0.200 0.917 0.205 0.801 0.167 0.709 −0.694 0.448

Note: Correlation coefficients (ρ) and the standard deviation (SDρ) as well as the slope and the standard deviation (SDslope) of the regression lines were calculated over all stimulation frequencies for each drug, concentration, and ROI and are presented as averages for each concentration and for each ROI type. The correlation coefficients were then analyzed for drug effect. *P < 0.05; **P < 0.005. In control conditions, the neuronal soma, astrocyte soma, and neuropil showed almost the same correlation strength. In neuropil and neurons, THIP displayed a concentration-dependent effect on the Ca2+−ΣLFP correlation. This correlation was also altered by Zolpidem administration at low concentrations, indicating that Zolpidem equally affected Ca2+ responses and ΣLFP. The strongest concentrations of both THIP and Zolpidem reduced the correlation and regression slopes between ΔCBF and ΣLFP in accordance with the decreased Ca2+ responses seen in Figures 4 and 5.

Figure 6.

Linear correlations between synaptic activity and the respective total Ca2+ responses. Graphs represent evoked Ca2+ responses as a function of the total synaptic activity (ΣLFP) before and after treatment with increasing concentrations of THIP (A, C, and E; yellow line: 1 μm; light green line: 10 μm; dark green line: 100 μm) or Zolpidem (B, D, and F; yellow line: 0.5 μm; orange line: 1 μm; red line: 10 μm). Postsynaptic activity as indicated by ΣLFP was correlated with Ca2+ activities in neuropil (A and B), neuronal soma (C and D), and astrocyte somas (E and F) under control conditions. Low concentrations of Zolpidem had no effect on the correlation between Ca2+ responses and ΣLFP in neuropil (B) and neuronal somas (D). The results of the correlation analyses are summarized in Table 5. Unlike Zolpidem, THIP decoupled Ca2+ responses in all cell types in a concentration-dependent manner.

Figure 6.

Linear correlations between synaptic activity and the respective total Ca2+ responses. Graphs represent evoked Ca2+ responses as a function of the total synaptic activity (ΣLFP) before and after treatment with increasing concentrations of THIP (A, C, and E; yellow line: 1 μm; light green line: 10 μm; dark green line: 100 μm) or Zolpidem (B, D, and F; yellow line: 0.5 μm; orange line: 1 μm; red line: 10 μm). Postsynaptic activity as indicated by ΣLFP was correlated with Ca2+ activities in neuropil (A and B), neuronal soma (C and D), and astrocyte somas (E and F) under control conditions. Low concentrations of Zolpidem had no effect on the correlation between Ca2+ responses and ΣLFP in neuropil (B) and neuronal somas (D). The results of the correlation analyses are summarized in Table 5. Unlike Zolpidem, THIP decoupled Ca2+ responses in all cell types in a concentration-dependent manner.

Zolpidem had a less obvious effect on the correlations between Ca2+ and ΣLFP. When analyzing all drug concentrations together, Zolpidem showed an overall significant effect on the correlation between Ca2+ responses in astrocyte soma and ΣLFP. This suggests that Zolpidem can uncouple the Ca2+ response in astrocyte soma and synaptic activity, similar to the effect of THIP. However, in contrast to THIP, Zolpidem had no overall effect on the correlation between Ca2+ response and ΣLFP in either neuropil or neuronal soma. Both the slopes of the correlation analysis and the correlation coefficients themselves were preserved during increases in synaptic GABAAR activity.

The observation that the Ca2+ signals in the 3 cellular compartments were linearly correlated to ΣLFP, as were ΔCBF and ΔCMRO2, suggests that the imaging and the electrophysiological, vascular, and metabolic data sets can be compared between experiments—that is, within and outside two-photon microscopy—when applying the same standard protocol. The data suggested that increases in Ca2+ correlated to increases in synaptic excitatory activity, whereas only increased tonic activity uncoupled Ca2+ activity in neuronal soma and neuropil from synaptic activity. On the other hand, both tonic and phasic inhibition uncoupled Ca2+ activity in astrocyte soma from synaptic activity.

## Discussion

The present study examined the influence of phasic and tonic inhibition on cortical activity by by-passing the local neuronal network. We have previously reported the feasibility of this approach in the cerebellar cortex and the somatosensory cortex, where we augmented the net inhibitory input by topical application of GABA agonists during stimulation (Caesar et al. 2003; Enager et al. 2009). It is difficult, if not impossible, to assess the impact of GABA receptor activity per se if the interneurons are not by-passed by means of specific drugs. In the present study, we employed Zolpidem and THIP as tool compounds, taking advantage of the functional differences in the mechanisms by which they each increase GABAergic inhibition. THIP application mimics an elevation of tonic inhibition in the brain, whereas Zolpidem increases the strength of the phasic response to synaptically released GABA during neurotransmission. Both phasic and tonic inhibition are important for maintaining a proper balance between excitation and inhibition in the brain, and disturbances in the GABAergic inhibitory system can result in a number of pathological conditions, such as epilepsy, schizophrenia, sleep disorders, Alzheimer's disease, and depression. Furthermore, tonic inhibition is important for controlling network dynamics in the brain, which is essential for higher cognitive functions and for regulating states of consciousness (Farrant and Nusser 2005; Brickley and Mody 2012).

Neurovascular coupling responses are driven by enhanced excitatory neurotransmission, with glutamate acting as a key player influencing the activity of astrocytes (Takano et al. 2006; Winship et al. 2007; Schummers et al. 2008; Nimmerjahn et al. 2009; Attwell et al. 2010; Girouard et al. 2010), excitatory and inhibitory neurons (Niessing et al. 2005; Enager et al. 2009; Lecrux et al. 2011), and pericytes (Peppiatt et al. 2006). Glutamate is believed to trigger increases in cytosolic Ca2+ that, via the activation of specific enzymatic pathways, lead to the release of vasoactive substances that control local blood flow. In comparison, activity-dependent energy use (metabolic coupling) is controlled by the activity of the Na+/K+-ATPase, which is mainly used to restore ionic gradients that are perturbed by synaptic and action potential currents (Creutzfeldt 1975; Erecinska and Silver 1989; Thompson et al. 2003; Viswanathan and Freeman 2007; Mathiesen et al. 2011; Lauritzen et al. 2012).

The present study used a pharmacological approach to assess the effects of synaptic and extrasynaptic GABAAR activities on neurovascular and metabolic coupling responses, linking activity-induced ΣLFP to ΔCBF, ΔCMRO2, and cytosolic Ca2+ transients in neurons and astrocytes. Our results demonstrated that increasing the gain of phasic synaptic GABAAR activity increased ΔCBF, ΔCMRO2, and ΣLFP at low Zolpidem concentrations, whereas high concentrations reduced ΔCBF and ΣLFP at high stimulation frequencies. We further found that increases in extrasynaptic GABAAR activity with THIP had a similar bidirectional effect on ΔCBF, ΔCMRO2, and ΣLFP. THIP also concentration-dependently suppressed cytosolic Ca2+ transients in neuronal and astrocyte soma and in neuropil, whereas Zolpidem increased Ca2+ transients in neuronal somas. Finally, all effects of synaptic and extrasynaptic GABAAR agonists on fluorescent Ca2+ signals, &triangle;CBF, and &triangle;CMRO2 were found to vary with the ΣLFP magnitude. Thus, our findings show that slight increases of the activity at both synaptic and extrasynaptic GABAARs enhanced responses at low stimulation frequencies, Whereas responses to intermediate- and high-frequency somatosensory afferent activity were suppressed. This is the first reported observation of GABAAR gating and amplification of responses at low stimulation frequencies.

### Impact of Zolpidem and THIP on Excitatory Neurotransmission

Activation of synaptic GABAAR results in the opening of a chloride conductance, leading to a fast inhibitory postsynaptic potential (IPSP) that provides robust suppression of neuronal output with fast temporal control. Zolpidem binds to the benzodiazepine-binding site of the receptor, leading to increased channel opening frequency, and has low affinity for receptors outside the synapse (Macdonald and Olsen 1994; Drasbek et al. 2007). Zolpidem is a positive allosteric modulator of synaptic GABAARs and requires synaptically released GABA to be able to increase synaptic GABAAR activity (Ebert and Wafford 2006). Zolpidem shows subtype selectivity, preferentially binding to the α1-subunit and requiring the presence of the GABAAR γ2-subunit (Macdonald and Olsen 1994; Drasbek et al. 2007). In layer V of the rat somatosensory cortex, Zolpidem increases the amplitude and charge of the IPSCs in fast-spiking (FS) interneurons, but not in low-threshold-spiking interneurons (Bacci, Rudolph et al. 2003). These data indicate that the GABAAR subunits are differently expressed in different types of interneurons. Hence, Zolpidem is assumed to be active only when GABA is released from presynaptic terminals of inhibitory neurons.

In contrast, at therapeutically relevant concentrations (i.e., in the micromolar range), in vivo THIP is specific for extrasynaptic GABAARs containing the δ-subunit (Krogsgaard-Larsen et al. 2004; Cremers and Ebert 2007). THIP is a superagonist of these receptors, leading to full activation upon binding, and studies in brain slices have confirmed THIP's specificity for extrasynaptic GABAAR (Drasbek and Jensen 2006; Drasbek et al. 2007; Vardya et al. 2008). GABAARs containing the δ-subunit have been identified in the mouse cerebral cortex (Drasbek et al. 2007; Krook-Magnuson et al. 2008), particularly in neurogliaform cells (Olah et al. 2009). Interneurons have been implicated in the neurovascular coupling responses, both in vivo and in vitro, by release of vasoactive substances and neuromodulators (Cauli et al. 2004; Enager et al. 2009; Karagiannis et al. 2009). Previous in vitro studies have shown that neuropeptide Y induces vasoconstriction, whereas vasoactive intestinal peptide (VIP) and nitric oxide cause vasodilation (Cauli et al. 2004; Olah et al. 2009). In vivo experiments in rats indicate that interneurons containing somatostatin (SOM), VIP, and parvalbumin (PV)—including neurogliaform cells—are activated during somatosensory stimulation (Enager et al. 2009; Lecrux et al. 2011). Neurogliaform cells abundantly express δ-subunit-containing GABAARs and are therefore targets for THIP. Nitric oxide release from these interneurons is expected to decrease in response to THIP, which is expected to directly affect the neurovascular response by reducing stimulation-induced vasodilation (Karagiannis et al. 2009). However, THIP's effect on interneurons is not general, since SOM-positive interneurons react only very weakly to low THIP concentrations (Drasbek and Jensen 2006; Drasbek et al. 2007; Vardya et al. 2008).

In line with the present study, a recent in vivo study found decreases in evoked neurovascular coupling responses after blocking GABAARs with picrotoxin in the rat somatosensory barrel cortex (Lecrux et al. 2011). This effect was ascribed to the suppressed inhibition of VIP/choline acetyltransferase interneurons on PV and calbindin interneurons. As PV and calbindin interneurons are the primary drivers of pyramidal neuron inhibition in layer 2/3, disinhibition of these interneurons results in increased inhibition of the pyramidal neurons and, consequently, reduced synaptic and hemodynamic responses. In line with this in vivo study, in vitro studies of cortical brain slices of postnatal day 14–42 have identified excitatory effects of THIP at low concentrations, which may also be explained by disinhibition (Drasbek and Jensen 2006; Drasbek et al. 2007; Vardya et al. 2008). Whole-cell patch-clamp recordings of cortical pyramidal neurons in layers 2/3 and 5 have shown increased firing frequency of spontaneous EPSPs and decreased spontaneous IPSPs in layer 2/3 neurons after application of 1 μm THIP (Drasbek and Jensen 2006), which has a greater effect on inhibitory than on excitatory neurons (Krook-Magnuson et al. 2008). These data indicate that low increases in tonic inhibition lead to increased network excitability due to suppression of interneuron activity.

In the present study, we report bidirectional effects of THIP, which can be best described as a suppression of cortical interneurons, leading to decreased inhibitory synaptic input and thereby increased network excitability at low concentrations, as well as an overall depressant effect at high concentrations when THIP binds to extrasynaptic GABAAR at both pyramidal cells and interneurons. The effects of Zolpidem on ΣLFP, &triangle;CBF, and ΔCMRO2 at low and high stimulation frequencies and concentrations, to some extent, parallel those of THIP. This may suggest that a common denominator for the effects of both THIP and Zolpidem on &triangle;CBF and &triangle;CMRO2 is modulation of the excitatory synaptic input via autapses on cortical inhibitory interneurons.

Both THIP and Zolpidem in the micromolar range show subtype selectivity in vivo, but increasing the concentration of either compound would possibly lead to activation of a more diverse population of GABAAR. Therefore, we expect non-specificity of Zolpidem and THIP for GABAAR subtypes at all but the lowest doses (Macdonald and Olsen 1994; Houston et al. 2012).

### Suppression of CBF and CMRO2 Responses by Increased GABAAR Activity

THIP and Zolpidem both increased ΣLFP amplitude at low concentrations, suggesting that a small increase in tonic and/or phasic inhibition leads to suppression of cortical interneurons (Vardya et al. 2008). The increased ΣLFP amplitude is reflected as increased ΔCBF and &triangle;CMRO2 at low concentrations and low stimulation frequencies. It is assumed that the increased ΣLFP is produced by increased transmembrane currents. At intermediate and high THIP concentrations, the ΣLFP was suppressed due to shunting inhibition (which increases with tonic conductance) as well as nonspecific effects of THIP on synaptic GABAARs (Farrant and Nusser 2005). Overall, the data for Zolpidem and THIP are in agreement with previous findings in rat somatosensory cortex and hippocampus. Radhakrishnan et al. (2011) reported that the magnitudes of ΔCBF and neural responses are suppressed by GABA and stimulation of ipsilateral somatosensory afferents (Berwick et al. 2004), and other studies have demonstrated that these responses are modulated by the activation of cellular elements downstream from the thalamocortical input rather than by the primary input alone (Franceschini et al. 2008; Angenstein et al. 2009; Harris et al. 2010). Our present data are also consistent with observations in rat cerebellum, showing that high concentrations of the GABAAR agonist muscimol attenuated ΔCBF and ΔCMRO2 responses (Mathiesen et al. 2011).

These findings could also be explained by direct effects on either blood vessels or astrocytes. An in vitro study of hippocampal microvessels and neocortex arterioles showed vasodilation of both GABA and the GABAAR agonist bicuculline, suggesting that GABAAR activation can directly modulate vessel diameter (Fergus and Lee 1997). We cannot rule out the possibility of a direct vasoactive effect of THIP and Zolpidem in our in vivo model; however, other in vivo findings currently do not support a direct effect of GABA on blood vessels (Lecrux et al. 2011).

Astrocytes express GABAARs in both hippocampal cultures and slice preparations (MacVicar et al. 1989), somatosensory cortical brain slices (Benedetti et al. 2011), and human brain cultures (Lee et al. 2011). However, in contrast to neuronal GABAARs, activation of astrocytic GABAARs causes an efflux of Cl into the extracellular space (MacVicar et al. 1989), where they may stabilize the Cl ion concentration in the neuronal synapses and thereby affect neuronal excitability (MacVicar et al. 1989; Fraser et al. 1995). Hence, the astrocyte GABAARs may contribute to interneuronal and/or neuronal inhibition by modulating the microenvironment (MacVicar et al. 1989; Fraser et al. 1995; Benedetti et al. 2011). Thus, the effects of Zolpidem and THIP reported in the present study may be explained partially by direct effects on astrocytic GABAARs, leading to efflux of Cl and a subsequent increase of intracellular astrocytic Ca2+. This in turn would induce release of vasodilators—mainly EETs and prostaglandins (Attwell et al. 2010; Lecrux et al. 2011)—ultimately leading to increases in cerebral blood flow and oxygen delivery. However, the specific subunit expression of GABAARs of cortical astrocytes in vivo remains to be identified, and more research is needed to fully understand the contribution of astrocytes to GABAergic inhibition in vivo.

### Effect of Tonic and Phasic Inhibition on Ca2+ Dynamics

Ca2+ is an intracellular messenger that has been proposed to mediate neurovascular coupling responses (Lauritzen 2005). Glutamate release during whisker pad stimulation increases intracellular Ca2+ in both neurons and astrocytes (Petzold and Murthy 2011). There is heavy debate over the involvement of astrocytes in the neurovascular coupling response. Several recent reports show divergent evidence of their possible role, some concluding that astrocytic Ca2+ responses are too slow to be likely contributors to neurovascular coupling responses (Nizar et al. 2013). However, we have recently demonstrated that cortical astrocytes in vivo do show fast transient Ca2+ responses on a timescale similar to that of neurons (Lind et al. 2013).

In the present study, Ca2+ transients in neuropil and neuronal and astrocyte somas were found to correlate with ΣLFP during control conditions. The Ca2+ activity in neuronal soma was significantly increased by the lowest Zolpidem concentration, corresponding to a small rise in phasic inhibition, whereas higher Zolpidem concentrations suppressed all responses. The enhanced Ca2+ response in neuronal soma at low Zolpidem concentrations may be due to increased interneuron-to-interneuron synaptic transmission (Hioki et al. 2013). Zolpidem contributes greatly to GABAAR activity and has strong effects on IPSCs in FS interneurons, which contain functional autapses for GABA (Bacci, Huguenard et al. 2003; Bacci, Rudolph et al. 2003). In turn, FS interneurons control the activity level of pyramidal cells in layer IV. Zolpidem is expected to increase suppression of FS interneurons via autapses (Bacci, Huguenard et al. 2003; Xu et al. 2013), leading to enhanced activity of layer IV SOM-positive interneurons (Xu et al. 2013). We suggest that this may explain why Zolpidem enhanced the Ca2+ response in neuronal somas at low concentrations, where its effect was specific for synaptic GABAARs, whereas high concentrations of Zolpidem decreased Ca2+ transients in all cell types. In comparison, all concentrations of THIP decreased Ca2+ responses in all ROIs, which is likely due to decreased excitability, as reported in the literature (Drasbek and Jensen 2006).

It is proposed that different types of interneurons are activated during whiskering in vivo (Gentet et al. 2010). During active whisking, non-FS interneurons expressing SOM, calretinin, or VIP dominate, whereas FS interneurons expressing PV are the primary source of inhibition during non-active whisking (Gentet et al. 2010; Gentet 2012). This would suggest that the observed enhanced Ca2+ response in neuronal somas would primarily occur through activation of non-FS interneurons, and not via FS as we have just proposed. However, in vivo studies have confirmed that activation of both FS and non-FS interneurons is involved in neurovascular coupling responses (Enager et al. 2009), and FS interneurons are activated during sensory stimulation (Gentet et al. 2012). Furthermore, FS PV-positive interneurons are one of the most abundant interneuronal cell types in layer 2/3, and exert feed-forward inhibition on pyramidal neurons (Freund and Katona 2007; Gentet et al. 2010; Gentet 2012). It has also been reported that gamma oscillations underlie hemodynamic responses (Niessing et al. 2005; Riera and Sumiyoshi 2010; Sumiyoshi et al. 2012), and gamma oscillations are induced and maintained by FS interneurons in the cortex (Cardin et al. 2009; Avermann et al. 2012). Together, these observations suggest that FS interneurons are indeed active during neurovascular coupling responses and may well be the effectors of Zolpidem action.

## Conclusion

Our data suggest that modulation of GABAAR activity has bidirectional effects on activity-dependent rises in hemodynamic and metabolic responses, which are explained by modulation of postsynaptic activity. The rise in hemodynamic responses during slight increases in GABAAR activity may be due to disinhibition of cortical interneurons, whereas the reduced hemodynamic responses during large increases in GABAAR activity may be caused by direct suppression of the overall excitability in all cell types. Our data suggest that low increments in activity at synaptic and extrasynaptic GABAARs amplify the vascular and metabolic responses to transient low-frequency somatosensory inputs and reduce the vascular and metabolic signals evoked by excitatory inputs at intermediate and higher frequencies. We argue that GABAAR gating of excitatory activity carried by somatosensory afferents appears to be a suitable mechanism for selective gating and amplification of sensory information mediated by low-frequency activity. The fact that the resulting rather complex activity pattern was found to influence ΔCBF and ΔCMRO2 suggests that these indirect markers of neural function can encode complex patterns of interaction between glutamatergic and GABAergic neurotransmission, which may be reflected in BOLD functional neuroimaging signals.

## Notes

We thank Micael Lønstrup for excellent technical assistance. We thank Krzysztof Kucharz for providing valuable feedback on the manuscript and excellent assistance with the figures and layout. We thank Jean-François Perrier for reading and commenting on the manuscript. We also thank Bjarke Ebert of Lundbeck A/S, Denmark, for kindly providing us with the pharmacological compound THIP. Conflict of Interest: None declared.

## References

Angenstein
F
Kammerer
E
Scheich
H
2009
.
The BOLD response in the rat hippocampus depends rather on local processing of signals than on the input or output activity. A combined functional MRI and electrophysiological study
.
J Neurosci
.
29
:
2428
2439
.
Armstrong-James
M
Fox
K
Das-Gupta
A
1992
.
Flow of excitation within rat barrel cortex on striking a single vibrissa
.
J Neurophysiol
.
68
:
1345
1358
.
Attwell
D
Buchan
AM
Charpak
S
Lauritzen
M
MacVicar
BA
Newman
EA
2010
.
Glial and neuronal control of brain blood flow
.
Nature
.
468
:
232
243
.
Attwell
D
Laughlin
SB
2001
.
An energy budget for signaling in the grey matter of the brain
.
J Cereb Blood Flow Metab
.
21
:
1133
1145
.
Austin
VC
Blamire
AM
Allers
KA
Sharp
T
Styles
P
Matthews
PM
Sibson
NR
2005
.
Confounding effects of anesthesia on functional activation in rodent brain: a study of halothane and alpha-chloralose anesthesia
.
Neuroimage
.
24
:
92
100
.
Avermann
M
Tomm
C
Mateo
C
Gerstner
W
Petersen
CC
2012
.
Microcircuits of excitatory and inhibitory neurons in layer 2/3 of mouse barrel cortex
.
J Neurophysiol
.
107
:
3116
3134
.
Bacci
A
Huguenard
JR
Prince
DA
2003
.
Functional autaptic neurotransmission in fast-spiking interneurons: a novel form of feedback inhibition in the neocortex
.
J Neurosci
.
23
:
859
866
.
Bacci
A
Rudolph
U
Huguenard
JR
Prince
DA
2003
.
Major differences in inhibitory synaptic transmission onto two neocortical interneuron subclasses
.
J Neurosci
.
23
:
9664
9674
.
Benedetti
B
Matyash
V
Kettenmann
H
2011
.
Astrocytes control GABAergic inhibition of neurons in the mouse barrel cortex
.
J Physiol
.
589
:
1159
1172
.
Berwick
J
Redgrave
P
Jones
M
Hewson-Stoate
N
Martindale
J
Johnston
D
Mayhew
JE
2004
.
Integration of neural responses originating from different regions of the cortical somatosensory map
.
Brain Res
.
1030
:
284
293
.
Bonvento
G
Charbonne
R
Correze
JL
Borredon
J
Seylaz
J
Lacombe
P
1994
.
Is alpha-chloralose plus halothane induction a suitable anesthetic regimen for cerebrovascular research?
Brain Res
.
665
:
213
221
.
Brickley
SG
Mody
I
2012
.
Extrasynaptic GABA(A) receptors: their function in the CNS and implications for disease
.
Neuron
.
73
:
23
34
.
Buzsaki
G
Kaila
K
Raichle
M
2007
.
Inhibition and brain work
.
Neuron
.
56
:
771
783
.
Caesar
K
Offenhauser
N
Lauritzen
M
2008
.
Gamma-aminobutyric acid modulates local brain oxygen consumption and blood flow in rat cerebellar cortex
.
J Cereb Blood Flow Metab
.
28
:
906
915
.
Caesar
K
Thomsen
K
Lauritzen
M
2003
.
Dissociation of spikes, synaptic activity, and activity-dependent increments in rat cerebellar blood flow by tonic synaptic inhibition
.
Proc Natl Acad Sci USA
.
100
:
16000
16005
.
Cardin
JA
Carlen
M
Meletis
K
Knoblich
U
Zhang
F
Deisseroth
K
Tsai
LH
Moore
CI
2009
.
Driving fast-spiking cells induces gamma rhythm and controls sensory responses
.
Nature
.
459
:
663
667
.
Cauli
B
Hamel
E
2010
.
Revisiting the role of neurons in neurovascular coupling
.
Front Neuroenergetics
.
2
:
9
.
Cauli
B
Tong
XK
Rancillac
A
Serluca
N
Lambolez
B
Rossier
J
Hamel
E
2004
.
Cortical GABA interneurons in neurovascular coupling: relays for subcortical vasoactive pathways
.
J Neurosci
.
24
:
8940
8949
.
Cremers
T
Ebert
B
2007
.
Plasma and CNS concentrations of Gaboxadol in rats following subcutaneous administration
.
Eur J Pharmacol
.
562
:
47
52
.
Creutzfeldt
O
1995
.
General neurophysiology of the cortex.
In:
Cortex Cerebri
.
Oxford
:
Oxford Science Publishers
. p.
131
163
.
Creutzfeldt
O
1975
.
Neurophysiological correlates of different functional states of the brain
. In:
Ingvar
DH
Lassen
NA
editors.
Brain Work. The coupling of Function, Metabolism and Blood Flow in the Brain
.
Copenhagen
:
Munksgaard
. p.
21
47
.
Drasbek
KR
Hoestgaard-Jensen
K
Jensen
K
2007
.
Modulation of extrasynaptic THIP conductances by GABAA-receptor modulators in mouse neocortex
.
J Neurophysiol
.
97
:
2293
2300
.
Drasbek
KR
Jensen
K
2006
.
THIP, a hypnotic and antinociceptive drug, enhances an extrasynaptic GABAA receptor-mediated conductance in mouse neocortex
.
Cereb Cortex
.
16
:
1134
1141
.
Ebert
B
Wafford
KA
2006
.
Benzodiazepine receptor agonists and insomnia: Is subtype selectivity lost in translation?
Drug Discov Today: Therapeutic Strategies
.
3
:
547
554
.
Enager
P
Piilgaard
H
Offenhauser
N
Kocharyan
A
Fernandes
P
Hamel
E
Lauritzen
M
2009
.
Pathway-specific variations in neurovascular and neurometabolic coupling in rat primary somatosensory cortex
.
J Cereb Blood Flow Metab
.
29
:
976
986
.
Erecinska
M
Silver
IA
1989
.
ATP and brain function
.
J Cereb Blood Flow Metab
.
9
:
2
19
.
Fabricius
M
Akgoren
N
Dirnagl
U
Lauritzen
M
1997
.
Laminar analysis of cerebral blood flow in cortex of rats by laser-Doppler flowmetry: a pilot study
.
J Cereb Blood Flow Metabol
.
17
:
1326
1336
.
Fabricius
M
Lauritzen
M
1996
.
Laser-Doppler evaluation of rat brain microcirculation: comparison with the [14C]-iodoantipyrine method suggests discordance during cerebral blood flow increases
.
J Cereb Blood Flow Metabol
.
16
:
156
161
.
Farrant
M
Nusser
Z
2005
.
Variations on an inhibitory theme: phasic and tonic activation of GABA(A) receptors
.
Nat Rev Neurosci
.
6
:
215
229
.
Fergus
A
Lee
KS
1997
.
GABAergic regulation of cerebral microvascular tone in the rat
.
J Cereb Blood Flow Metab
.
17
:
992
1003
.
Franceschini
MA
Nissila
I
Wu
W
Diamond
SG
Bonmassar
G
Boas
DA
2008
.
Coupling between somatosensory evoked potentials and hemodynamic response in the rat
.
Neuroimage
.
41
:
189
203
.
Fraser
DD
Duffy
S
Angelides
KJ
Perez-Velazquez
JL
Kettenmann
H
MacVicar
BA
1995
.
GABAA/benzodiazepine receptors in acutely isolated hippocampal astrocytes
.
J Neurosci
.
15
:
2720
2732
.
Freund
TF
Katona
I
2007
.
Perisomatic inhibition
.
Neuron
.
56
:
33
42
.
Garrett
KM
Gan
J
1998
.
Enhancement of gamma-aminobutyric acid A receptor activity by alpha-chloralose
.
J Pharmacol Exp Ther
.
285
:
680
686
.
Gentet
LJ
2012
.
Functional diversity of supragranular GABAergic neurons in the barrel cortex
.
Front Neural Circuits
.
6
:
52
.
Gentet
LJ
Avermann
M
Matyas
F
Staiger
JF
Petersen
CC
2010
.
Membrane potential dynamics of GABAergic neurons in the barrel cortex of behaving mice
.
Neuron
.
65
:
422
435
.
Gentet
LJ
Kremer
Y
Taniguchi
H
Huang
ZJ
Staiger
JF
Petersen
CC
2012
.
Unique functional properties of somatostatin-expressing GABAergic neurons in mouse barrel cortex
.
Nat Neurosci
.
15
:
607
612
.
Girouard
H
Bonev
Hannah
RM
Meredith
A
Aldrich
RW
Nelson
MT
2010
.
Astrocytic endfoot Ca2+ and BK channels determine both arteriolar dilation and constriction
.
Proc Natl Acad Sci USA
.
107
:
3811
3816
.
Gjedde
A
Johannsen
P
Cold
GE
Ostergaard
L
2005
.
Cerebral metabolic response to low blood flow: possible role of cytochrome oxidase inhibition
.
J Cereb Blood Flow Metabol
.
25
:
1183
1196
.
Harris
S
Jones
M
Zheng
Y
Berwick
J
2010
.
Does neural input or processing play a greater role in the magnitude of neuroimaging signals?
Front Neuroenergetics
.
2
:
1
7
.
Harrison
TC
Sigler
A
Murphy
TH
2009
.
Simple and cost-effective hardware and software for functional brain mapping using intrinsic optical signal imaging
.
J Neurosci Methods
.
182
:
211
218
.
Hildebrandt
IJ
Su
H
Weber
WA
2008
.
Anesthesia and other considerations for in vivo imaging of small animals
.
ILAR J
.
49
:
17
26
.
Hioki
H
Okamoto
S
Konno
M
Kameda
H
Sohn
J
Kuramoto
E
Fujiyama
F
Kaneko
T
2013
.
Cell type-specific inhibitory inputs to dendritic and somatic compartments of parvalbumin-expressing neocortical interneuron
.
J Neurosci
.
33
:
544
555
.
Houston
CM
McGee
TP
Mackenzie
G
Troyano-Cuturi
K
Rodriguez
PM
Kutsarova
E
Diamanti
E
Hosie
AM
Franks
NP
Brickley
SG
2012
.
Are extrasynaptic GABAA receptors important targets for sedative/hypnotic drugs?
J Neurosci
.
32
:
3887
3897
.
C
Nedergaard
M
2007
.
Glial regulation of the cerebral microvasculature
.
Nat Neurosci
.
10
:
1369
1376
.
Jacob
TC
Moss
SJ
Jurd
R
2008
.
GABA(A) receptor trafficking and its role in the dynamic modulation of neuronal inhibition
.
Nat Rev Neurosci
.
9
:
331
343
.
Jespersen
SN
Ostergaard
L
2012
.
The roles of cerebral blood flow, capillary transit time heterogeneity, and oxygen tension in brain oxygenation and metabolism
.
J Cereb Blood Flow Metab
.
32
:
264
277
.
Karagiannis
A
Gallopin
T
David
C
Battaglia
D
Geoffroy
H
Rossier
J
Hillman
EM
Staiger
JF
Cauli
B
2009
.
Classification of NPY-expressing neocortical interneurons
.
J Neurosci
.
29
:
3642
3659
.
Kocharyan
A
Fernandes
P
Tong
XK
Vaucher
E
Hamel
E
2008
.
Specific subtypes of cortical GABA interneurons contribute to the neurovascular coupling response to basal forebrain stimulation
.
J Cereb Blood Flow Metab
.
28
:
221
231
.
Koehler
RC
Roman
RJ
Harder
DR
2009
.
Astrocytes and the regulation of cerebral blood flow
.
Trends Neurosci
.
32
:
160
169
.
Krogsgaard-Larsen
P
Frolund
B
Liljefors
T
Ebert
B
2004
.
GABA(A) agonists and partial agonists: THIP (Gaboxadol) as a non-opioid analgesic and a novel type of hypnotic
.
Biochem Pharmacol
.
68
:
1573
1580
.
Krogsgaard-Larsen
P
Johnston
GA
1978
.
Structure-activity studies on the inhibition of GABA binding to rat brain membranes by muscimol and related compounds
.
J Neurochem
.
30
:
1377
1382
.
Krook-Magnuson
EI
Li
P
Paluszkiewicz
SM
Huntsman
MM
2008
.
Tonically active inhibition selectively controls feedforward circuits in mouse barrel cortex
.
J Neurophysiol
.
100
:
932
944
.
Lauritzen
M
2005
.
Opinion: Reading vascular changes in brain imaging: is dendritic calcium the key?
Nat Rev Neurosci
.
6
:
77
85
.
Lauritzen
M
Mathiesen
C
Schaefer
K
Thomsen
KJ
2012
.
Neuronal inhibition and excitation, and the dichotomic control of brain hemodynamic and oxygen responses
.
Neuroimage
.
62
:
1040
1050
.
Lecrux
C
Toussay
X
Kocharyan
A
Fernandes
P
Neupane
S
Levesque
M
Plaisier
F
Shmuel
A
Cauli
B
Hamel
E
2011
.
Pyramidal neurons are “neurogenic hubs” in the neurovascular coupling response to whisker stimulation
.
J Neurosci
.
31
:
9836
9847
.
Lee
M
Schwab
C
McGeer
PL
2011
.
Astrocytes are GABAergic cells that modulate microglial activity
.
Glia
.
59
:
152
165
.
Lind
BL
Brazhe
AR
Jessen
SB
Tan
FC
Lauritzen
MJ
2013
.
Rapid stimulus-evoked astrocyte Ca2+ elevations and hemodynamic responses in mouse somatosensory cortex in vivo
.
Proc Natl Acad Sci USA
.
110
:
E4678
E4687
.
Lindauer
U
Villringer
A
Dirnagl
U
1993
.
Characterization of CBF response to somatosensory stimulation—model and influence of anesthetics
.
Am J Physiol
.
264
:
H1223
H1228
.
Macdonald
RL
Olsen
RW
1994
.
GABAA receptor channels
.
Annu Rev Neurosci
.
17
:
569
602
.
MacVicar
BA
Tse
FW
Crichton
SA
Kettenmann
H
1989
.
GABA-activated Cl-channels in astrocytes of hippocampal slices
.
J Neurosci
.
9
:
3577
3583
.
Mathiesen
C
Caesar
K
Lauritzen
M
2000
.
Temporal coupling between neuronal activity and blood flow in rat cerebellar cortex as indicated by field potential analysis
.
J Physiol (Lond)
.
523
:
235
246
.
Mathiesen
C
Caesar
K
Thomsen
K
Hoogland
TM
Witgen
BM
Brazhe
A
Lauritzen
M
2011
.
Activity-dependent increases in local oxygen consumption correlate with postsynaptic currents in the mouse cerebellum in vivo
.
J Neurosci
.
31
:
18327
18337
.
Nielsen
A
Lauritzen
M
2001
.
Coupling and uncoupling of activity-dependent increases of neuronal activity and blood flow in rat somatosensory cortex
.
J Physiol
.
533
:
773
785
.
Niessing
J
Ebisch
B
Schmidt
KE
Niessing
M
Singer
W
Galuske
RAW
2005
.
Hemodynamic signals correlate tightly with synchronized gamma oscillations
.
Science
.
309
:
948
951
.
Nimmerjahn
A
Kirchhoff
F
Kerr
JND
Helmchen
F
2004
.
Sulforhodamine 101 as a specific marker of astroglia in the neocortex in vivo
.
Nat Methods
.
1
:
31
37
.
Nimmerjahn
A
Mukamel
EA
Schnitzer
MJ
2009
.
Motor behavior activates Bergmann glial networks
.
Neuron
.
62
:
400
412
.
Nizar
K
Uhlirova
H
Tian
P
Saisan
PA
Cheng
Q
Reznichenko
L
Weldy
KL
Steed
TC
Sridhar
VB
MacDonald
CL
et al
2013
.
In vivo stimulus-induced vasodilation occurs without IP3 receptor activation and may precede astrocytic calcium increase
.
J Neurosci
.
33
:
8411
8422
.
Olah
S
Fule
M
Komlosi
G
Varga
C
Baldi
R
Barzo
P
Tamas
G
2009
.
Regulation of cortical microcircuits by unitary GABA-mediated volume transmission
.
Nature
.
461
:
1278
1281
.
Peppiatt
CM
Howarth
C
Mobbs
P
Attwell
D
2006
.
Bidirectional control of CNS capillary diameter by pericytes
.
Nature
.
443
:
700
704
.
Petzold
GC
Murthy
VN
2011
.
Role of astrocytes in neurovascular coupling
.
Neuron
.
71
:
782
797
.
H
Wu
W
Boas
D
Franceschini
MA
2011
.
Study of neurovascular coupling by modulating neuronal activity with GABA
.
Brain Res
.
1372
:
1
12
.
Raichle
ME
Mintun
MA
2006
.
Brain work and brain imaging
.
Annu Rev Neurosci
.
29
:
449
476
.
Riera
JJ
Sumiyoshi
A
2010
.
Brain oscillations: ideal scenery to understand the neurovascular coupling
.
Curr Opin Neurol
.
23
:
374
381
.
Schummers
J
Yu
H
Sur
M
2008
.
Tuned responses of astrocytes and their influence on hemodynamic signals in the visual cortex
.
Science
.
320
:
1638
1643
.
Semyanov
A
Walker
MC
Kullmann
DM
Silver
RA
2004
.
Tonically active GABA(A) receptors: modulating gain and maintaining the tone
.
Trends Neurosci
.
27
:
262
269
.
Stosiek
C
Garaschuk
O
Holthoff
K
Konnerth
A
2003
.
In vivo two-photon calcium imaging of neuronal networks
.
Proc Natl Acad Sci USA
.
100
:
7319
7324
.
Sumiyoshi
A
Suzuki
H
Ogawa
T
Riera
JJ
Shimokawa
H
Kawashima
R
2012
.
Coupling between gamma oscillation and fMRI signal in the rat somatosensory cortex: its dependence on systemic physiological parameters
.
Neuroimage
.
60
:
738
746
.
Takano
T
Tian
GF
Peng
WG
Lou
NH
Libionka
W
Han
XN
Nedergaard
M
2006
.
Astrocyte-mediated control of cerebral blood flow
.
Nat Neurosci
.
9
:
260
267
.
Thompson
JK
Peterson
MR
Freeman
RD
2003
.
Single-neuron activity and tissue oxygenation in the cerebral cortex
.
Science
.
299
:
1070
1072
.
Vardya
I
Drasbek
KR
Dosa
Z
Jensen
K
2008
.
Cell type-specific GABA A receptor-mediated tonic inhibition in mouse neocortex
.
J Neurophysiol
.
100
:
526
532
.
Viswanathan
A
Freeman
RD
2007
.
Neurometabolic coupling in cerebral cortex reflects synaptic more than spiking activity
.
Nat Neurosci
.
10
:
1308
1312
.
Winship
IR
Plaa
N
Murphy
TH
2007
.
Rapid astrocyte calcium signals correlate with neuronal activity and onset of the hemodynamic response in vivo
.
J Neurosci
.
27
:
6268
6272
.
Xu
H
Jeong
HY
Tremblay
R
Rudy
B
2013
.
Neocortical somatostatin-expressing GABAergic interneurons disinhibit the thalamorecipient layer 4
.
Neuron
.
77
:
155
167
.
Zhu
XH
Zhang
Y
Tian
RX
Lei
H
Zhang
NY
Zhang
XL
Merkle
H
Ugurbil
K
Chen
W
2002
.
Development of O-17 NMR approach for fast imaging of cerebral metabolic rate of oxygen in rat brain at high field
.
Proc Natl Acad Sci USA
.
99
:
13194
13199
.