The state of somatosensory cortex during neuromodulation
Abstract
During behavioral quiescence, such as slow-wave sleep and anesthesia, the neocortex is in a deactivated state characterized by the presence of slow oscillations. During arousal, slow oscillations are absent and the neocortex is in an activated state that greatly impacts information processing. Neuromodulators acting in neocortex are believed to mediate these state changes, but the mechanisms are poorly understood. We investigated the actions of noradrenergic and cholinergic activation on slow oscillations, cellular excitability, and synaptic inputs in thalamocortical slices of somatosensory cortex. The results show that neuromodulation abolishes slow oscillations, dampens the excitability of principal cells, and rebalances excitatory and inhibitory synaptic inputs in thalamocortical-recipient layers IV–III. Sensory cortex is much more selective about the inputs that can drive it. The source of neuromodulation is critically important in determining this selectivity. Cholinergic activation suppresses the excitatory and inhibitory conductances driven by thalamocortical and intracortical inputs. Noradrenergic activation suppresses the excitatory conductance driven by intracortical inputs but not by thalamocortical inputs and enhances the inhibitory conductance driven by thalamocortical inputs but not by intracortical inputs. Thus noradrenergic activation emphasizes thalamocortical (sensory) inputs relative to intracortical inputs, while cholinergic activation suppresses both.
the thalamus and neocortex generate a variety of electrical activities during different behavioral states that have profound consequences on signals flowing through them. During slow-wave sleep and anesthesia, cortical networks are typically in a so-called deactivated or synchronized state that consists of spontaneous “slow oscillations,” which are characterized by rhythmic cycles of synaptically mediated depolarization and increased firing (Up states) followed by a decrease of synaptic inputs leading to membrane hyperpolarization and cessation of firing (Down states) (Cowan and Wilson 1994; Steriade et al. 1993b). During arousal, vigilance, and paradoxical sleep, cortical networks are typically in a so-called activated or desynchronized state characterized by the absence of slow oscillations (Castro-Alamancos 2009; Moruzzi and Magoun 1949). In sensory cortex, thalamocortical responses driven by sensory signals are very significantly transformed by activated states during arousal (Castro-Alamancos 2002b, 2004a, 2004b; Castro-Alamancos and Oldford 2002; Hirata and Castro-Alamancos 2011; Stoelzel et al. 2009), but the underlying mechanisms of these changes are poorly understood.
Cortical activation is readily induced in sleeping/anesthetized animals in vivo by waking or by stimulating the brain stem reticular formation or the basal forebrain. This effect is thought to be due to the cortical actions of neuromodulators such as norepinephrine (NE) and acetylcholine (Castro-Alamancos 2004b; Constantinople and Bruno 2011; Metherate et al. 1992; Saper et al. 2010; Steriade et al. 1993a; Vanderwolf 1988). Indeed, cholinergic neurons within brain stem and basal forebrain nuclei discharge vigorously during paradoxical sleep and wakefulness (el Mansari et al. 1989; Jones 2008; Steriade et al. 1990), while noradrenergic neurons in the locus coeruleus discharge robustly during high levels of vigilance and attention (Aston-Jones and Bloom 1981; Foote et al. 1980; Hobson et al. 1975). Thus noradrenergic and cholinergic neurons appear to discharge robustly when slow oscillations are absent, but the effects of these neuromodulators on slow oscillations are poorly understood. For instance, the effects of these neuromodulators on the network activity observed in cortical slices have not been studied before.
Using thalamocortical slices from adult mice that produce persistent network activity resembling slow oscillations, we studied the impact of noradrenergic and cholinergic activation on network activity and on thalamocortical and intracortical synaptic inputs in the somatosensory cortex. Our investigation focused on principal excitatory (regular-spiking pyramidal and spiny stellate) cells in the main thalamocortical-recipient layers (IV–III) of somatosensory cortex; a few identified inhibitory [fast spiking (FS), based on firing and morphology] cells were also studied. We found that neuromodulation sets an activated state that consists of dampened excitability in principal cells, abolishment of slow oscillations, and rebalancing of synaptic inputs, with an emphasis on thalamocortical inputs during noradrenergic activation.
METHODS
All procedures were reviewed and approved by the Animal Care Committee of Drexel University. Slices were prepared as previously described (Rigas and Castro-Alamancos 2007, 2009) from adult (>8 wk) CD-1 mice. Mice were deeply anesthetized with an overdose of ketamine hydrochloride. Upon losing all responsiveness to a strong tail pinch the animal was decapitated and the brain was rapidly extracted. Slices (400 μm thick) were cut in the thalamocortical plane (Agmon and Connors 1991) with a vibratome. Slices were transferred to an interface chamber where they were bathed constantly (1–1.5 ml/min) with artificial cerebrospinal fluid (ACSF) at 32.5°C. The ACSF contained (in mM) 126 NaCl, 3.5 KCl, 1.25 NaH2Po4, 26 NaHCO3, 1 MgSO4·7H2O, 10 dextrose, and 1 CaCl2·2H2O. This buffer results in the generation of persistent network activity resembling slow oscillations in cortical slices (McCormick et al. 2003). Field potential (FP) recordings were made in the area of the barrel cortex with low-impedance (∼0.5 MΩ) glass pipettes filled with ACSF. Blind whole cell recordings were obtained mostly from cells in layers IV and III (IV–III) of somatosensory cortex with patch electrodes of 4- to 12-MΩ impedance. A couple of cells were located in the lower part of layer II (at the border with layer III), and a few were layer V pyramidal cells whose apical dendrites were patched in layer IV. For current-clamp recordings, the electrodes were filled with internal solution containing (in mM) 135 K-gluconate, 4 KCl, 2 NaCl, 0.2 EGTA, 10 Tris-phosphocreatine, 0.3 Tris-GTP, 10 HEPES, and 4 MgATP (290 mosM). Under our conditions, the Nernst equilibrium potential is −81 mV for Cl− and −96.7 mV for K+. In most cases, the internal solution contained neurobiotin (0.2%) to label the recorded cells. Junction potentials were corrected.
After each experiment, the slices were fixated in 4% paraformaldehyde with 1% glutaraldehyde, later cryoprotected with sucrose (30%), and resectioned on a cryostat (80 μm). Sections were incubated in 0.3% hydrogen peroxide followed by 0.2% Triton X-100 and by incubation in 2% goat serum. Incubation with ABC reagent (Vector Labs) occurred overnight. The following day, diaminobenzidine was applied to the sections. After color development, sections were mounted and cleared in xylene. The labeled cells were faithfully reconstructed with Neurolucida (Microbrightfield).
Concentric bipolar stimulating electrodes were used to electrically stimulate the thalamus (thalamocortical) and cortex (intracortical). The FP electrode was first used to identify the cortical region with the strongest and shortest latency response evoked by thalamic stimulation. Intracellular recordings were obtained from layers IV–III adjacent to the FP electrode. Thalamocortical excitatory postsynaptic potentials (EPSPs) met two criteria: they depressed at 20 Hz and had short latencies (<2.5 ms). The intracortical stimulating electrode was placed lateral to the recording electrodes (∼400 μm) in layer III. Drugs were dissolved in ACSF at the indicated concentrations.
In these slices, during the control period the membrane potential (Vm) of cortical cells fluctuates between Up and Down states, and Up states are well known to affect both the intrinsic and synaptic properties of cortical cells (McCormick et al. 2003; Rigas and Castro-Alamancos 2009; Waters and Helmchen 2006). Therefore, all measurements during the control period were performed during the Down state, i.e., all traces used were determined to be free of Up states for at least 200 ms before the onset of the stimulus tested. Up states were typically detected off-line by using a threshold detector in the FP recording (see Rigas and Castro-Alamancos 2007, 2009). Relatively rare transient FP events that are not Up states are easily rejected by setting the detection algorithm to reject short-duration events (<50 ms). In addition, all detected events are sent to a sorting algorithm (similar to those used to sort spikes), and this allows classification of all detected Up states on the basis of several projections (e.g., principal components, etc). Finally, all detected events (selected or unselected as Up states) are inspected by eye to ensure that the procedure was adequate.
Intracellular current pulses, extracellular electrical stimulation, and/or glutamate pulses were applied (alternated) at a minimum of 5 s between each other, and each recurred at an interval of 10 s or higher. Negative current pulses (100 ms; 0.1–0.2 nA) were applied at least every 10 s to monitor input resistance. Synaptic potentials were evoked at least 5 s apart from each electrode, and during each stimulus the Vm could be set at a different value by applying negative and positive current pulses (>100 ms before the synaptic stimulus onset), up to the level that produced spontaneous firing. This allowed derivation of the reversal potential for each point of the synaptic response (Moore and Nelson 1998) and estimation of the excitatory and inhibitory synaptic conductance (Gsynexc and Gsyninh, 0 mV and −81 mV reversal potential, respectively) before and during neuromodulation (Shu et al. 2003). Reversal potential for each point was estimated by calculating the y-intercept of the best fit line between the baseline-corrected (subtracting values 5 ms prior to the stimulus) Vm values (x-axis) and the Vm values (y-axis). Total conductance for each point was estimated by calculating the inverse slope of the best fit line between the injected current (x-axis) and the Vm values (y-axis). The synaptic conductance (Gsyn) is the total conductance minus the baseline conductance. Gsyninh for each point was estimated as the product of the reversal potential and Gsyn divided by the Nernst equilibrium potential for inhibitory (Cl− is −81 mV) currents. Gsynexc for each point was Gsyn minus Gsyninh. We assume that membrane capacitance is a constant for each cell during our experiments and that during the period we measure (<15 ms after stimulus) the synaptic conductance is composed primarily of glutamatergic excitation and Cl−-mediated inhibition.
Glutamate iontophoresis was performed by applying negative current pulses (∼100 nA; 500–1,000 ms) to glass micropipettes filled with glutamate (1 M in saline; 7.4 pH). The micropipettes targeted the dendrites and were usually located in layer III ∼100–300 μm away from the recording electrode. To study isolated inhibitory postsynaptic potentials (IPSPs), synaptic responses were evoked by an intracortical stimulating electrode in the presence of CNQX (10–20 μM) and AP5 (25–50 μM). Variable current pulses were tested during control and during neuromodulation by applying short sequences of 500-ms current pulses of different intensities every 2 s (usually between −0.5 and +0.4 nA at 0.1- or 0.05-nA intervals; repeated several times). Population measurements were done during 10 min before (control) and 5–20 min after (neuromodulation) drug application, when the effects of the drug were stable.
For statistical analyses, data were first tested for normality with the Shapiro-Wilk test. If the data were considered normally distributed, parametric statistics were applied (repeated-measures ANOVA or paired t-test). Otherwise, we applied nonparametric statistics (Wilcoxon signed ranks for paired comparisons, Mann-Whitney for nonpaired comparisons, Kruskal-Wallis for multiple groups).
RESULTS
Neuromodulators abolish slow oscillations.
We conducted simultaneous intracellular (whole cell) and FP recordings from microelectrodes placed in layers IV–III of the somatosensory cortex, as in previous studies (Rigas and Castro-Alamancos 2007, 2009). Figure 1A shows spontaneous slow oscillations consisting of Up and Down states recorded from a layer IV pyramidal cell (cell 9 is reconstructed in Fig. 5A). In the intracellular recording, the Up state is characterized by barrages of postsynaptic potentials (PSPs) that depolarize the recorded neuron by 5–15 mV for the duration of the network event. The Up state is clearly reflected in the FP recording as a synchronous population (network) event. During the Down state, cortical cells are relatively hyperpolarized and there is little synaptic or network activity. Figure 1B shows that electrically stimulating the thalamus triggers Up states in neocortex very effectively. Thus, in thalamocortical-recipient cortical cells, the thalamocortical-evoked intracellular and FP responses consist of a short-latency PSP (<5 ms) that is followed by a longer-latency (>10 ms) Up state (Rigas and Castro-Alamancos 2007). Non-thalamocortical-recipient cells only produce the longer-latency Up state (see below).
![Fig. 1. Fig. 1.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930001.jpeg)
Fig. 1.Noradrenergic activation abolishes spontaneous and evoked Up states. A: spontaneous intracellular (whole cell) and field potential (FP) activity showing slow oscillations in an identified somatosensory cortex cell (cell 9 in Fig. 5A) during control and NE (50 μM). Vm, membrane potential. B: thalamocortical-evoked Up states in the same cell shown in A during control and NE. C: population data showing the incidence of spontaneous Up states and the FP amplitude (>10 ms) of thalamocortical-evoked Up states during control and during 2 different doses of NE (10 and 50 μM). *P < 0.01.
Bath application of NE (50 μM) completely abolished spontaneous (Fig. 1, A and C) and thalamocortical-evoked Up states but not the short-latency PSP that precedes the Up state (Fig. 1B). This effect was evident in both the FP and intracellular recordings. Population analyses revealed that NE (10–50 μM) abolished spontaneous Up states in all experiments tested (n = 22 at 50 μM and n = 15 at 10 μM; P < 0.01); the incidence of spontaneous Up states was nil during NE (Fig. 1C). To estimate the effect of NE on thalamocortical-evoked Up states, we measured the peak amplitude of the long-latency FP response (15–50 ms after stimulus; Rigas and Castro-Alamancos 2007) and found a significant decrease for both doses (n = 13 at 50 μM and n = 15 at 10 μM; P < 0.01). However, only the higher dose (50 μM) produced a complete abolishment of evoked Up states in all experiments (Fig. 1C). Therefore, we used this dose for the remainder of the study.
We next set out to determine the effect of cholinergic activation on spontaneous and evoked Up states by testing the effects of the cholinergic agonist carbachol (CA). Figure 2, A and B, show examples of thalamocortical-evoked Up states during control and after CA (5 μM) application in two different cells. Just like NE, CA application abolished spontaneous and thalamocortical evoked Up states but not the short-latency PSP. Population analyses revealed a robust suppression of spontaneous (n = 10 at 5 μM and n = 11 at 1 μM; P < 0.01) and thalamocortical-evoked (n = 12 at 5 μM and n = 11 at 1 μM; P < 0.01) Up states in all experiments (Fig. 2C). However, a complete abolishment of spontaneous and thalamocortical-evoked Up states was observed in all experiments only with the higher dose (5 μM) of CA (Fig. 2C). Therefore, we used this dose for the remainder of the study.
![Fig. 2. Fig. 2.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930002.jpeg)
Fig. 2.Cholinergic activation abolishes spontaneous and evoked Up states. A: thalamocortical-evoked Up states in a thalamocortical-recipient cell during control and carbachol (CA, 5 μM). B: thalamocortical-evoked Up states in a non-thalamocortical-recipient cell (cell 31 in Fig. 5B) during control and CA (5 μM). C: population data showing the incidence of spontaneous Up states and the FP amplitude (>10 ms) of thalamocortical-evoked Up states during control and during 2 different doses of CA (1 and 5 μM). *P < 0.01.
It is worth noting that the cell in Fig. 2B corresponds to an identified layer V pyramidal cell (cell 31 is reconstructed in Fig. 5B) patched in layer IV. This non-thalamocortical-recipient cell responded to thalamocortical stimulation with an Up state but with a negligible short-latency PSP, indicating a sparse thalamocortical input. Importantly, the FP shows a clear short-latency population response, reflecting a significant short-latency response in other thalamocortical-recipient cells. Thus, by triggering Up states, thalamocortical activity is capable of driving (albeit at long latencies) cells that receive rather poor thalamocortical input, but neuromodulators abolish this long-latency activity, resulting in a more restricted short-latency drive of thalamocortical-recipient cells in neocortex during neuromodulation.
Specific receptors mediate the actions of neuromodulators on slow oscillations.
We next tested the effects of specific receptor agonists on spontaneous (Fig. 3, A and B) and thalamocortical-evoked (Fig. 3C) Up states. We employed doses found previously to have significant actions in the somatosensory thalamus and cortex under similar experimental conditions (Castro-Alamancos 2002a; Castro-Alamancos and Calcagnotto 2001; Gil et al. 1997). Regarding noradrenergic activation, we found that the α1-adrenergic agonist phenylephrine (50 μM) completely abolished spontaneous and thalamocortical-evoked Up states. However, the α2-adrenergic agonist clonidine (100 μM) significantly suppressed but did not completely abolish the occurrence of spontaneous Up states and did not significantly suppress thalamocortical-evoked Up states. Similarly, the β-adrenergic agonist isoproterenol (50 μM) significantly suppressed, but did not completely abolish, spontaneous and thalamocortical-evoked Up states.
![Fig. 3. Fig. 3.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930003.jpeg)
Fig. 3.Effects of specific noradrenergic and cholinergic agonists on spontaneous and evoked Up states. A: mean effect over time since drug application on the incidence of spontaneous Up states per minute. B and C: population data showing the effects of agonists on the incidence of spontaneous Up states (B) and on the FP amplitude (>10 ms) of thalamocortical-evoked Up states (C). *P < 0.01.
Regarding cholinergic activation, since CA is a robust muscarinic agonist (like acetylcholine, it has a higher affinity for muscarinic receptors than for nicotinic receptors), we tested the effect of nicotine (10 μM) and found that it did not suppress spontaneous (Fig. 3, A and B) or thalamocortical-evoked (Fig. 3C) Up states. In some experiments nicotine increased the incidence of spontaneous Up states, but this was not significant overall. In conclusion, activation of muscarinic or α1-adrenergic receptors completely abolishes spontaneous slow oscillations and thalamocortical-evoked Up states in neocortex. Activation of β-adrenergic receptors also robustly suppresses Up states.
Neuromodulators suppress the input resistance of cortical cells.
The abolishment of spontaneous and evoked Up states produced by NE was accompanied on average by a small (1.72 ± 0.5 mV) but significant depolarization of Vm (−68.7 ± 0.8 vs. −66.9 ± 1 mV; P < 0.01; n = 40 cells; Fig. 4A). NE also caused on average a significant reduction in input resistance monitored with two different methods. The first method involved measuring the voltage drop caused by a negative current pulse (100 ms; −0.2 nA) and resulted in a significant decrease (15.6%) in input resistance (57.7 ± 3.6 vs. 48.5 ± 3.7 MΩ; P < 0.01, n = 40 cells; Fig. 4B). The second method involved measuring the decay time constant of the voltage after the offset of the 100-ms current pulse. This method is obviously less sensitive to changes in access resistance that may confound changes in input resistance. It also resulted in a significant decrease (1.75 ± 0.2 ms or 19.3 ± 2.9%) in input resistance (9.05 ± 0.5 vs. 7.3 ± 0.4 ms; P < 0.01, n = 40 cells; Fig. 4C). Figure 4D depicts for each cell the change in input resistance (decay time constant) as a function of the change in Vm and reveals that there is barely a significant linear relationship between these two variables (red trace in Fig. 4D), so that cells that depolarize the most tend to show less change in input resistance. This may be explained by the fact that depolarization tends to increase input resistance in many cortical cells (Rigas and Castro-Alamancos 2009; Waters and Helmchen 2006), compensating the decrease caused by NE. Moreover, 22 of the 40 cells tested with NE were labeled with neurobiotin and reconstructed and are shown in Fig. 5A (cell numbers in Fig. 5A correspond to those in Fig. 4D). Most cells in layers IV–III of somatosensory cortex are excitatory pyramidal or spiny stellate cells (Feldmeyer et al. 2002; Schubert et al. 2003), and there is also a small portion (∼10%) of FS inhibitory cells (Gibson et al. 1999; Kruglikov and Rudy 2008). Indeed, most of the cells we recorded were pyramidal or spiny stellate cells located in layers IV–III. In addition, there were several cells that had their cell bodies located in layer V (cells 20–22). These cells correspond to dendritic patches because the recording electrodes were located in layers IV–III (verified with photos of electrode placements taken during each experiment). Removing these dendritic recordings from the population analyses did not change statistical significance or any conclusions. A comparison of identified cells located in layer IV (n = 9) to those located in layer III (n = 9) did not result in a significant difference between them regarding the effects of NE on either Vm (P = 0.9) or input resistance (P = 0.8). In addition, two of the cells located in layer IV (cell 11 in Fig. 5A and Fig. 7B, inset) had a multipolar morphology with expansive dendrites and electrophysiologically corresponded to an FS cell. These cells were considered separately.
![Fig. 4. Fig. 4.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930004.jpeg)
Fig. 4.Neuromodulators affect Vm and input resistance (Rin) of cortical cells. The effect of NE on Vm (A) and Rin was estimated by measuring the voltage drop caused by a (−0.2 nA; 100 ms) current pulse (B) or by measuring the decay time constant (τ) after pulse offset (C). The change in Rin caused by NE is plotted as a function of the change in Vm (D). Results are also shown for the effects of CA (E–H). The numbers in D and H correspond to the cells reconstructed in Fig. 5. *P < 0.01.
![Fig. 5. Fig. 5.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930005.jpeg)
Fig. 5.Identification of some of the cells studied. Cells are grouped based on whether they were tested for the effects of NE (A) or CA (B) during control conditions or during isolated inhibitory postsynaptic potentials (IPSPs) (C).
The abolishment of spontaneous and evoked Up states produced by CA was accompanied on average by a small (−0.9 ± 0.5 mV) but significant hyperpolarization of Vm (−68.4 ± 1.1 vs. −69.4 ± 1.1 mV; P < 0.05, n = 18 cells; Fig. 4E) that was more highly significant if two identified pyramidal cells located in upper layer III or layer II (cells 23 and 24 in Fig. 5B) were removed from the population (n = 16; P < 0.01). CA also caused on average a significant reduction in input resistance measured with the voltage drop method (61.9 ± 5.3 vs. 45.6 ± 4.9 MΩ; P < 0.0, n = 18 cells; Fig. 4F) and the decay time constant method (8.09 ± 0.9 vs. 6.1 ± 0.5 ms; P < 0.01, n = 18 cells; Fig. 4G). Figure 4H depicts for each cell the change in input resistance (decay time constant) as a function of the change in Vm and reveals that there is no significant linear relationship between these two variables (red trace in Fig. 4H). Moreover, 9 of the 18 cells tested with CA were labeled with neurobiotin and reconstructed and are shown in Fig. 5B (cell numbers in Fig. 5B correspond to those in Fig. 4H). As per the NE cells, most of the CA cells were pyramidal or spiny stellate cells located in layers IV–III. Notably, the effects of cholinergic activation on Vm hyperpolarization reported here resemble those recently reported by another study that investigated its actions in rat somatosensory cortex (Eggermann and Feldmeyer 2009).
Neuromodulators affect the excitability of cortical cells.
To further determine the effects of NE and CA on the intrinsic excitability of cortical cells, we conducted two sets of tests on cellular excitability. The first involved testing variable negative and positive current pulses injected in each cell before and during neuromodulation. The second involved testing the effects of glutamate pulses and a constant positive current pulse on each cell before and after neuromodulation.
In the first set of cells, we tested current pulses (500 ms) of varying amplitudes (from −0.5 to +0.1 nA at 0.1-nA intervals) on the sustained Vm change measured just before the offset of the pulse. Current pulses were applied during the Down state in control and during neuromodulation. We also determined the effects of the neuromodulators on the ability of the same positive current pulses (several pulses between +0.1 and +0.4 nA) to evoke spikes during control and during neuromodulation. The majority of our cells displayed inward rectification (i.e., input resistance increased with depolarization around resting membrane potential). This means that, in many cortical cells, slope conductance is not linear but negative, and this can be explained because of activation of voltage-dependent noninactivating inward currents during subthreshold depolarization that sum with outward currents (Connors et al. 1982; Spain et al. 1987; Stafstrom et al. 1982). Figure 6, A and B, show the responses of a spiny stellate cell (cell 13 in Fig. 5A) and a pyramidal cell (cell 8 in Fig. 5A) to current pulses before and during NE. There were two major effects of neuromodulation that were similar for both NE and CA. First, input resistance decreased significantly for all current pulses during NE (n = 18 cells; Fig. 6C) or CA (n = 20 cells; Fig. 6D). Second, the number of spikes evoked by the same positive current pulses (mean firing rate evoked by several increasing pulses) was significantly reduced for virtually every cell during NE (Fig. 6E; n = 22, P < 0.01) and CA (Fig. 6F; n = 11, P < 0.01); during neuromodulation more current is required to evoke the same number of spikes than during control. There were a few notable exceptions. During NE, the identified FS cells (cell 11 in Fig. 5A and the cell in Fig. 7B) strongly increased their firing rate in response to the same current pulses compared with control (cell 11 changed from 45 ± 11 Hz during control to 178 ± 7 Hz during NE). Also during NE, an identified stellate cell (cell 10 in Fig. 5A) also increased its firing rate (from 16 ± 2 Hz during control to 25 ± 2 Hz during NE). During CA, the only cell showing an increase in evoked firing was cell 31, which is an identified layer V pyramidal cell patched in layer IV.
![Fig. 6. Fig. 6.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930006.jpeg)
Fig. 6.Effects of neuromodulators on cellular excitability. A and B: examples of responses to variable negative and positive intracellular current pulses from 2 identified cells (shown in Fig. 5A) before and during NE. C and D: population data showing the effects of NE (C) or CA (D) on Rin measured before current offset from variable current pulses (500 ms). x-Axis shows the steady-state Vm during the current pulse (indicated within symbols). D: equivalent data for the effects of CA. E and F: spikes evoked by the same positive current pulses before and during NE (E) or CA (F). *P < 0.01. EK, Nernst equilibrium potential for K+.
In the second set of cells, we tested the effect of glutamate iontophoretic pulses (1,000 ms) applied to the dendrites of the recorded cells. Each glutamate pulse alternated with a positive intracellular current pulse (500 ms) applied through the recording electrode to monitor evoked spikes. This allowed measurement of responses evoked by both intracellular current and glutamate pulses before and during neuromodulation. Figure 7 shows examples from four different cells (Fig. 7, A–D; 5 trials overlaid) and population data (Fig. 7, E and F). We measured both the area of depolarization produced by the glutamate pulses (Fig. 7E) and the number of spikes evoked by the intracellular current pulses (Fig. 7F). During NE, the depolarization evoked by glutamate pulses and the number of spikes evoked by either the glutamate or current pulses were significantly suppressed (Fig. 7A). This was the case in all the cells tested (n = 11), except for an identified FS cell (partially reconstructed multipolar cell shown in Fig. 7B, inset). This was the only cell in which NE increased the number of spikes evoked by both glutamate and current pulses (similar to the effect of NE on current pulses in cell 11 described above). Thus in principal cells (regular-spiking excitatory cells), NE produced a significant suppression of responses to both glutamate and intracellular current pulses (n = 10, P < 0.01; Fig. 7, E and F). CA (Fig. 7C) produced a similar effect (n = 7, P < 0.01; Fig. 7, E and F). In contrast, nicotine (Fig. 7D), which does not block Up states, had no significant effects on responses to either glutamate or intracellular current pulses (n = 4; Fig. 7, E and F). We also tested the effect of TTX in several cells and found that, while abolishing ongoing synaptic events and evoked spikes, it had little effect on the constant direct depolarization produced by the glutamate pulse (Fig. 7D, gray traces).
![Fig. 7. Fig. 7.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930007.jpeg)
Fig. 7.Effects of neuromodulators on responses evoked by glutamate pulses. A–D: examples of responses to the same glutamate iontophoretic pulses (1,000 ms) and to the same positive intracellular current pulses (500 ms) before and during NE (A and B), CA (C), or nicotine (D). Each panel overlays 5 trials. Insets in A, C, and D show negative current pulse used to monitor Rin. Also depicted is the mean ± SE firing rate evoked by the intracellular current pulse during control and during neuromodulation. Inset in B shows the recorded cell reconstructed. D overlays gray traces after application of TTX (1 μM). E: population data showing the Vm area measured from baseline and triggered by the glutamate pulses before and during NE, CA, or nicotine. F: population data for the cells in E, showing the number of spikes triggered by the intracellular current pulse before and during NE, CA, or nicotine. *P < 0.01.
In conclusion, neuromodulation caused by NE or CA suppresses the input resistance of cortical cells, resulting in a high-conductance state during which the excitability of principal (excitatory) cells in sensory cortex is suppressed. However, as shown in two identified cells, NE increases the excitability of FS (inhibitory) cells. Thus the abolishment of Up states in sensory cortex by neuromodulators is accompanied by increases in membrane conductance that appear to shunt the ability of excitatory currents to drive action potentials in principal cells.
Neuromodulators rebalance synaptic inputs.
In a group of cells, we measured the effects of NE and CA on short-latency thalamocortical and intracortical evoked responses. Synaptic responses were evoked by a single stimulus to the thalamus or locally in the upper layers of cortex, as previously described (Rigas and Castro-Alamancos 2007, 2009). Figure 8A shows the effect of NE on a spiny stellate cell in layer IV (cell 13) and on the simultaneously recorded FP response. NE caused a significant increase in the amplitude of the short-latency FP response; at the same time, the cell was slightly depolarized and the EPSP showed only weak changes. To further assess the impact of neuromodulators on these responses, we recorded the synaptic responses of each cell at different Vm set by variable current pulses (Fig. 8B). This was then used to estimate the Gsynexc and Gsyninh evoked by thalamocortical and intracortical stimuli (Fig. 8C). Thus Fig. 8, D–G, show population data measuring the amplitude of FP responses (Fig. 8D), the slope of the EPSP measured at resting membrane potential (Fig. 8E), and the amplitude of Gsynexc (Fig. 8F) and Gsyninh (Fig. 8G) evoked by thalamocortical and intracortical stimuli during control and neuromodulation. Thalamocortical responses had onset latencies <2.3 ms (1.9 ± 0.2 ms) and fast-rising EPSPs that depressed with frequency. Intracortical responses had onset latencies <3 ms (2.7 ± 0.2) that were slower rising. All measurements were performed during the short-latency rising phases of the evoked responses (<6 ms) in order to minimize the effect of evoked Up states and inhibitory conductance on excitatory responses.
The results show that NE significantly enhanced the amplitude of FP thalamocortical responses (n = 11, P < 0.01) but suppressed intracortical responses (n = 22, P < 0.01). The enhanced FP thalamocortical response caused by NE did not reflect a significant change in either EPSP slope (n = 11) or Gsynexc (n = 10). The suppressed FP intracortical response caused by NE reflected a significant reduction both in EPSP slope (n = 22) and in Gsynexc (n = 12). CA significantly suppressed the amplitude of FP thalamocortical (n = 13, P < 0.01) and intracortical (n = 10, P < 0.01) responses. The suppressed FP thalamocortical response caused by CA reflected a significant reduction in EPSP slope (n = 13) and in Gsynexc (n = 9). The suppressed FP intracortical response caused by CA reflected a significant reduction in EPSP slope (n = 16) and in Gsynexc (n = 12).
Interestingly, NE significantly enhanced the inhibitory conductance evoked by thalamocortical stimuli but not by intracortical stimuli (Fig. 8G). In contrast, CA significantly suppressed the inhibitory conductance evoked by both thalamocortical and intracortical stimuli. To further determine the impact of these neuromodulators on inhibitory responses, in another group of cells (some of which are reconstructed in Fig. 5C), we measured the effects of NE and CA on isolated IPSPs evoked by a local intracortical stimulating electrode in the presence of CNQX (20 μM) and AP5 (50 μM). Figure 9A shows Gsyninh for two identified cells (cells 33 and 38 in Fig. 5C) before and during application of NE. Although NE may have some effects on long-latency inhibitory responses, the short-latency inhibitory response (<15 ms), which is mediated by GABAA receptors (not shown), is not significantly suppressed by NE (n = 6; Fig. 9C). In contrast, CA had a significant suppressive effect on inhibitory responses, similar to previous work (Kruglikov and Rudy 2008). Figure 9B shows Gsyninh for two identified cells (cell 36 in Fig. 5C and a partially reconstructed spiny stellate cell, not shown) before and during application of CA. Inhibitory synaptic responses were significantly suppressed by CA (n = 7, P < 0.01; Fig. 9C).
![Fig. 8. Fig. 8.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930008.jpeg)
Fig. 8.Effects of neuromodulators on synaptic responses. A: example showing the effect of NE on FP thalamocortical responses and postsynaptic potentials recorded at resting membrane potential from an identified cell (cell 13 in Fig. 5A). B: thalamocortical-evoked postsynaptic potentials recorded from the same cell at different Vm set by intracellular current injection before and during NE for the cell in A. C: excitatory (Gsynexc) and inhibitory (Gsyninh) synaptic conductance derived from the potentials in B. D–G: population data showing the effects of NE and CA on the short-latency (<5 ms) FP amplitude (D), excitatory postsynaptic potential (EPSP) slope (E), and Gsynexc (F) and Gsyninh (G) measured from thalamocortical and intracortical (stimulating electrode in layer III) responses. *P < 0.01.
![Fig. 9. Fig. 9.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930009.jpeg)
Fig. 9.Effects of neuromodulators on isolated inhibitory synaptic responses. A and B: Gsyninh evoked by intracortical stimulation in 4 different cells during CNQX and AP5 before and during application of NE (A) or CA (B). C: population data showing the peak amplitude of Gsyninh before and during NE or CA. *P < 0.01.
Our results, obtained from identified FS cells, imply that FS cells are more excitable during NE. If this is the case, they may spontaneously spike more during NE. Consequently, the Vm noise measured in principal cells during isolated IPSPs should reflect this increase. We measured the spontaneous Vm noise by computing the FFT power spectrum of Vm in several cells (only spontaneous subthreshold activity periods without firing or evoked responses were measured) at 5–100 Hz frequency, which reflects mainly synaptic activity (Jacobson et al. 2005). The results show that this frequency range significantly increased during NE in four of the six principal cells tested (n = 6), while Vm noise did not increase in any of the principal cells subjected to CA (n = 7). This supports the contention that FS cells are more excitable during NE, but not during CA.
These results indicate that neuromodulators impact synaptic inputs differently. Noradrenergic activation suppresses intracortical but not thalamocortical excitatory inputs and enhances inhibitory inputs driven by thalamocortical activity. The effect of NE on inhibitory responses is likely caused by enhanced excitability of FS cells driven by thalamocortical inputs, and not by an effect on inhibitory synapses. Cholinergic activation suppresses excitatory and inhibitory synaptic inputs driven by either thalamocortical or intracortical activity. The main difference between the effects of these two neuromodulators in somatosensory cortex lies in the major influence of thalamocortical excitatory and inhibitory inputs during noradrenergic activation compared with cholinergic activation.
Neuromodulators do not abolish abnormal network oscillations.
The suppression of the slowly inactivating K+ current ID with application of a low dose (25 μM) of 4-AP (Storm 1990) transforms normal slow oscillations in somatosensory cortex into abnormal hypersynchronous and hyperexcitable oscillations (Castro-Alamancos and Tawara-Hirata 2007). Since application of NE and CA completely abolished normal slow oscillations, we wanted to determine whether these neuromodulators would also abolish abnormal oscillations. Figure 10 shows the typical activity produced in somatosensory cortex during 4-AP (25 μM). The spontaneous activity is fairly regular and consists of long paroxysmal depolarizing shifts (long events) that recur every few minutes (Fig. 10A). The depolarization produced by each long event lasts many seconds (5–15 s) and drives ∼10-Hz oscillations on top of it (Fig. 10B; see Castro-Alamancos et al. 2007; Castro-Alamancos and Tawara-Hirata 2007). The long events are followed by much shorter paroxysmal depolarizing shifts (short events) that recur every few seconds (∼5 s; based on interevent time histograms). The depolarization produced by each short event lasts between a few hundred milliseconds and a few seconds (<3 s) and can also drive short epochs of ∼10-Hz oscillations on top of it (Fig. 10C). When the period of short events stops, it gives rise to a variable period of silence during which the synaptic (subthreshold) noise increases until a new long event occurs and the sequence is repeated.
![Fig. 10. Fig. 10.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930010.jpeg)
Fig. 10.Typical spontaneous abnormal network activity observed in somatosensory cortex during 4-AP (25 μM). A: simultaneous whole cell and FP recordings in somatosensory cortex showing activity consisting of 2 different types of discharges: long (B) and short (C) events. The typical pattern consists of a prominent paroxysmal depolarizing shift (long event), which decays slowly back to baseline and as it does it produces multiple smaller discharges (short events). Once the short events stop, there is a period of quiescence until a new long event occurs. D: effect of CA (5 μM) for the experiment in A.
We found that application of either NE or CA did not abolish ongoing abnormal network activity (Fig. 11). Instead, these modulators tended to slightly decrease the strength of the long events. To quantify these effects we detected, sorted, and measured short and long events separately during equivalent time periods (15 min) before and during NE or CA application. Each long event was measured during a 15-s time window from onset, while each short event was measured during a 3-s time window.
![Fig. 11. Fig. 11.](/cms/10.1152/jn.00256.2012/asset/images/medium/z9k0151214930011.jpeg)
Fig. 11.Neuromodulators do not abolish abnormal network activity. Effect of NE and CA on long and short discharges caused by 4-AP. The incidence of long and short events per minute (A), the firing rate during the long and short events (B), the area of depolarization from baseline (C), and the FFT power of the FP activity (D and E) are shown. Measurements of firing rate, Vm area, and FFT power were done during 15 s and 3 s for long and short events, respectively. *P < 0.01.
Regarding long events, we found that either NE (n = 12) or CA (n = 10) did not affect the incidence of long events (Fig. 11A) but both neuromodulators decreased the number of spikes (Fig. 11B; P < 0.05) as well as the area of depolarization from baseline (Fig. 11C; P < 0.01) that each long event produced. Moreover, a power spectrum analysis of the FP activity during long events showed that both NE and CA significantly suppressed the 5–25 Hz range (Fig. 11, D and E; P < 0.05), which corresponds to the range of ∼10-Hz oscillations that occur on top of the long events (Castro-Alamancos et al. 2007; Castro-Alamancos and Tawara-Hirata 2007).
Regarding short events, NE increased the incidence of short events (Fig. 11A; P < 0.05) and increased the area of depolarization (Fig. 11C; P < 0.05), while CA had no significant effects on these measures. Moreover, while NE had no significant effect on the FP power spectrum of short events (Fig. 11D), CA significantly suppressed FP activity of short events in the 5–50 Hz range (Fig. 11E; 5–10 Hz, 20–30 Hz, and 35–50 Hz ranges; P < 0.05). These results indicate that while NE and CA completely abolish normal slow oscillations, they do not abolish abnormal oscillations in the same network. Instead, the main effect of neuromodulation is to suppress the strength of slowly recurring long paroxysmal depolarizing shifts (long events).
DISCUSSION
We found that noradrenergic and cholinergic neuromodulation in somatosensory cortex produce a higher conductance state in principal cells of layers IV–III and a suppression of intracortical excitatory synaptic inputs. The result is a less excitable network that abolishes normal persistent network activity (slow oscillations) typical of quiescent behavioral states (slow-wave sleep) but not the abnormal network activity typical of pathological states (seizures). In addition, neuromodulators rebalance synaptic inputs in different ways. Noradrenergic activation facilitates excitatory and inhibitory thalamocortical (sensory) inputs relative to intracortical inputs, while cholinergic activation dampens both thalamocortical and intracortical synaptic inputs. Thus the state of sensory cortical networks is greatly transformed by neuromodulation; it is much more selective about the inputs that can drive it. Moreover, the source of neuromodulation is critically important so that noradrenergic activation emphasizes thalamocortical (sensory) inputs relative to intracortical inputs.
Neuromodulation and persistent network activity.
During quiescence, cortical networks generate slow oscillations consisting of Up and Down states. It is well known that Up states impact the excitability of cortical cells (McCormick et al. 2003; Rigas and Castro-Alamancos 2009; Waters and Helmchen 2006). An important question is whether Up states are akin to the activated states produced by neuromodulation. If this was the case, then neuromodulators would simply depolarize cells to the level of the Up state Vm. Instead, we find that neuromodulators abolish the generation of Up states and can either slightly depolarize Vm (NE) or hyperpolarize it (CA). Even the slight depolarization produced by NE is not equivalent to the much larger depolarization produced by Up states. Thus cholinergic and noradrenergic neuromodulation do not transform Down states into a steady-state Up state; instead, they abolish slow oscillations. Moreover, while Vm of cells is closer to the Down state during cholinergic or noradrenergic activation, they clearly differ from this state because of the higher conductance and reduced excitability.
Many studies have previously investigated the effects of various neuromodulators on cortical cells in vitro (e.g., Eggermann and Feldmeyer 2009; Gil et al. 1997; Gulledge and Stuart 2005; Kawaguchi and Shindou 1998; Kruglikov and Rudy 2008; McCormick 1992; Xiang et al. 1998) but not in adult animals and under conditions that resemble the deactivated (slow oscillation) state in vivo. Similar to a previous study (Eggermann and Feldmeyer 2009), we found that cholinergic activation hyperpolarizes many principal cells (spiny stellate and pyramidal) in layers IV–III. Although it is important to point out that the previous studies were done with submerged slices, young rats, and 2 mM Ca2+ ACSF as opposed to an interface chamber, adult mice, and 1 mM Ca2+ ACSF as used here. We also found that noradrenergic activation depolarizes these principal cells. Despite the differential effects on Vm between cholinergic and noradrenergic activation, in both cases principal cells were less excitable during neuromodulation. In contrast to principal cells, FS cells were more excitable during noradrenergic activation, an effect that has been observed in rat frontal cortex (Kawaguchi and Shindou 1998). The resulting activated state produced by neuromodulators consists in a higher conductance state in principal cells during which inputs have more difficulty in driving them, likely because of shunting. The dampened excitability observed in principal cells together with the suppression of intracortical synaptic efficacy likely underlies the abolishment of slow oscillations during neuromodulation.
Neuromodulation and synaptic inputs.
Thalamocortical activity triggers short-latency synaptic responses in thalamocortical-recipient cells and longer-latency Up states in most cortical cells (Rigas and Castro-Alamancos 2007). Our results show that neuromodulation affects very differently the short-latency thalamocortical synaptic responses and the longer-latency responses driven by Up states. By triggering Up states, thalamocortical activity is capable of driving cells at long latencies that receive rather poor thalamocortical input, but neuromodulators abolish this long-latency activity, resulting in a more restricted short-latency drive of thalamocortical-recipient cells during neuromodulation. Consequently, during neuromodulation cortical cells are much more selective about the synaptic inputs that can drive them. Indeed, a suppression and focusing of sensory (thalamocortical) responses in somatosensory cortex during cortical activation has been demonstrated repeatedly in both anesthetized and behaving animals (Castro-Alamancos 2002b, 2004a, 2004b; Castro-Alamancos and Oldford 2002; Hirata and Castro-Alamancos 2011).
The impact of cholinergic neuromodulators on short-latency intracortical and thalamocortical synaptic responses have been studied previously, and our results agree with previous studies showing that muscarinic receptor activation suppresses both thalamocortical and intracortical synaptic inputs (Eggermann and Feldmeyer 2009; Gil et al. 1997; Oldford and Castro-Alamancos 2003); although activation of nicotinic receptors enhances thalamocortical synaptic efficacy, this requires direct application of nicotine, likely because of the lower affinity of acetylcholine and CA for nicotinic compared with muscarinic receptors. In contrast, the effects of noradrenergic activation on thalamocortical and intracortical synaptic inputs have been less studied previously. We found that while intracortical excitatory inputs were suppressed by noradrenergic activation, thalamocortical excitatory inputs were not. In fact, thalamocortical synapses appear to be able to compensate the dampening of cellular excitability observed in principal cells during noradrenergic activation.
Because FS cells mediate recurrent inhibition driven by thalamocortical activity (Gibson et al. 1999; Sun et al. 2006) and FS cells increase their excitability during noradrenergic activation (Kawaguchi and Shindou 1998), but not during cholinergic activation (Kawaguchi 1997; Xiang et al. 1998), we should expect an increase in thalamocortical-evoked inhibitory responses during this state. Indeed, inhibitory responses driven by thalamocortical activity were enhanced by noradrenergic activation, while those driven by intracortical activity were not. The enhanced inhibitory responses driven by thalamocortical activity appear to be caused by the enhanced excitability of FS cells and cannot be attributed to a change in the efficacy of inhibitory synapses, which did not change.
The state of sensory cortex during neuromodulation.
The state of sensory cortex has several common features during noradrenergic and cholinergic activation. First, there is a suppression of excitability in principal cells. In general, principal cells are more selective about the inputs that can drive them. Second, the efficacy of intracortical synaptic connections is suppressed. Third, slow oscillations, which define the deactivated state, are abolished. The first two features surely contribute to the third feature.
There are also several differences that set different cortical modes during cholinergic and noradrenergic activation. In fact, in the somatosensory thalamus these neuromodulators set very distinct signal-to-noise ratios for sensory transmission, which are much larger during noradrenergic activation (Hirata et al. 2006). In neocortex, the excitability of FS cells seems to be enhanced selectively by noradrenergic activation (Kawaguchi 1997; Kawaguchi and Shindou 1998; Xiang et al. 1998). The enhanced excitability of FS cells may result in the differential generation of high-frequency gamma rhythms during these states (Cardin et al. 2009). Another important feature that is different between noradrenergic and cholinergic activation in neocortex is that while the efficacy of thalamocortical inputs is suppressed by cholinergic activation, it is not suppressed by noradrenergic activation. In fact, both excitatory and inhibitory responses driven by thalamocortical activity are enhanced during noradrenergic activation, relative to intracortical responses. The fact that noradrenergic activation emphasizes transmission from the sensory periphery, both in the somatosensory thalamus and neocortex, correlates well with the fact that this state occurs during vigilance and focused attention (Aston-Jones et al. 1994).
GRANTS
This work was supported by the National Institutes of Health.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the author(s).
AUTHOR CONTRIBUTIONS
Author contributions: M.F., G.V., and M.A.C.-A. performed experiments; M.F. and M.A.C.-A. analyzed data; M.F. and M.A.C.-A. interpreted results of experiments; M.F. and M.A.C.-A. edited and revised manuscript; M.F. and M.A.C.-A. approved final version of manuscript; M.A.C.-A. conception and design of research; M.A.C.-A. prepared figures; M.A.C.-A. drafted manuscript.
REFERENCES
- Thalamocortical responses of mouse somatosensory (barrel) cortex in vitro. Neuroscience 41: 365–379, 1991.
Crossref | PubMed | Web of Science | Google Scholar . - Nonrepinephrine-containing locus coeruleus neurons in behaving rats exhibit pronounced responses to non-noxious environmental stimuli. J Neurosci 1: 887–900, 1981.
Crossref | PubMed | Web of Science | Google Scholar . - Locus coeruleus neurons in monkey are selectively activated by attended cues in a vigilance task. J Neurosci 14: 4467–4480, 1994.
Crossref | PubMed | Web of Science | Google Scholar . - Driving fast-spiking cells induces gamma rhythm and controls sensory responses. Nature 459: 663–667, 2009.
Crossref | PubMed | Web of Science | Google Scholar . - Absence of rapid sensory adaptation in neocortex during information processing states. Neuron 41: 455–464, 2004a.
Crossref | PubMed | Web of Science | Google Scholar . - Cortical up and activated states: implications for sensory information processing. Neuroscientist 15: 625–634, 2009.
Crossref | PubMed | Web of Science | Google Scholar . - Dynamics of sensory thalamocortical synaptic networks during information processing states. Prog Neurobiol 74: 213–247, 2004b.
Crossref | PubMed | Web of Science | Google Scholar . - Properties of primary sensory (lemniscal) synapses in the ventrobasal thalamus and the relay of high-frequency sensory inputs. J Neurophysiol 87: 946–953, 2002a.
Link | Web of Science | Google Scholar . - Role of thalamocortical sensory suppression during arousal: focusing sensory inputs in neocortex. J Neurosci 22: 9651–9655, 2002b.
Crossref | PubMed | Web of Science | Google Scholar . - High-pass filtering of corticothalamic activity by neuromodulators released in the thalamus during arousal: in vitro and in vivo. J Neurophysiol 85: 1489–1497, 2001.
Link | Web of Science | Google Scholar . - Cortical sensory suppression during arousal is due to the activity-dependent depression of thalamocortical synapses. J Physiol 541: 319–331, 2002.
Crossref | PubMed | Web of Science | Google Scholar . - Resonance (approximately 10 Hz) of excitatory networks in motor cortex: effects of voltage-dependent ion channel blockers. J Physiol 578: 173–191, 2007.
Crossref | PubMed | Web of Science | Google Scholar . - Area-specific resonance of excitatory networks in neocortex: control by outward currents. Epilepsia 48: 1572–1584, 2007.
Crossref | PubMed | Web of Science | Google Scholar . - Electrophysiological properties of neocortical neurons in vitro. J Neurophysiol 48: 1302–1320, 1982.
Link | Web of Science | Google Scholar . - Effects and mechanisms of wakefulness on local cortical networks. Neuron 69: 1061–1068, 2011.
Crossref | PubMed | Web of Science | Google Scholar . - Spontaneous firing patterns and axonal projections of single corticostriatal neurons in the rat medial agranular cortex. J Neurophysiol 71: 17–32, 1994.
Link | Web of Science | Google Scholar . - Cholinergic filtering in the recurrent excitatory microcircuit of cortical layer 4. Proc Natl Acad Sci USA 106: 11753–11758, 2009.
Crossref | PubMed | Web of Science | Google Scholar . - Unitary characteristics of presumptive cholinergic tegmental neurons during the sleep-waking cycle in freely moving cats. Exp Brain Res 76: 519–529, 1989.
Crossref | PubMed | Web of Science | Google Scholar . - Synaptic connections between layer 4 spiny neurone-layer 2/3 pyramidal cell pairs in juvenile rat barrel cortex: physiology and anatomy of interlaminar signalling within a cortical column. J Physiol 538: 803–822, 2002.
Crossref | PubMed | Web of Science | Google Scholar . - Impulse activity of locus coeruleus neurons in awake rats and monkeys is a function of sensory stimulation and arousal. Proc Natl Acad Sci USA 77: 3033–3037, 1980.
Crossref | PubMed | Web of Science | Google Scholar . - Two networks of electrically coupled inhibitory neurons in neocortex. Nature 402: 75–79, 1999.
Crossref | PubMed | Web of Science | Google Scholar . - Differential regulation of neocortical synapses by neuromodulators and activity. Neuron 19: 679–686, 1997.
Crossref | PubMed | Web of Science | Google Scholar . - Cholinergic inhibition of neocortical pyramidal neurons. J Neurosci 25: 10308–10320, 2005.
Crossref | PubMed | Web of Science | Google Scholar . - Noradrenergic activation amplifies bottom-up and top-down signal-to-noise ratios in sensory thalamus. J Neurosci 26: 4426–4436, 2006.
Crossref | PubMed | Web of Science | Google Scholar . - Effects of cortical activation on sensory responses in barrel cortex. J Neurophysiol 105: 1495–1505, 2011.
Link | Web of Science | Google Scholar . - Sleep cycle oscillation: reciprocal discharge by two brainstem neuronal groups. Science 189: 55–58, 1975.
Crossref | PubMed | Web of Science | Google Scholar . - Subthreshold voltage noise of rat neocortical pyramidal neurones. J Physiol 564: 145–160, 2005.
Crossref | PubMed | Web of Science | Google Scholar . - Modulation of cortical activation and behavioral arousal by cholinergic and orexinergic systems. Ann NY Acad Sci 1129: 26–34, 2008.
Crossref | PubMed | Web of Science | Google Scholar . - Selective cholinergic modulation of cortical GABAergic cell subtypes. J Neurophysiol 78: 1743–1747, 1997.
Link | Web of Science | Google Scholar . - Noradrenergic excitation and inhibition of GABAergic cell types in rat frontal cortex. J Neurosci 18: 6963–6976, 1998.
Crossref | PubMed | Web of Science | Google Scholar . - Perisomatic GABA release and thalamocortical integration onto neocortical excitatory cells are regulated by neuromodulators. Neuron 58: 911–924, 2008.
Crossref | PubMed | Web of Science | Google Scholar . - Neurotransmitter actions in the thalamus and cerebral cortex and their role in neuromodulation of thalamocortical activity. Prog Neurobiol 39: 337–388, 1992.
Crossref | PubMed | Web of Science | Google Scholar . - Persistent cortical activity: mechanisms of generation and effects on neuronal excitability. Cereb Cortex 13: 1219–1231, 2003.
Crossref | PubMed | Web of Science | Google Scholar . - Cellular bases of neocortical activation: modulation of neural oscillations by the nucleus basalis and endogenous acetylcholine. J Neurosci 12: 4701–4711, 1992.
Crossref | PubMed | Web of Science | Google Scholar . - Spatio-temporal subthreshold receptive fields in the vibrissa representation of rat primary somatosensory cortex. J Neurophysiol 80: 2882–2892, 1998.
Link | Web of Science | Google Scholar . - Brain stem reticular formation and activation of the EEG. Electroencephalogr Clin Neurophysiol 1: 455–473, 1949.
Crossref | PubMed | Google Scholar . - Input-specific effects of acetylcholine on sensory and intracortical evoked responses in the “barrel cortex” in vivo. Neuroscience 117: 769–778, 2003.
Crossref | PubMed | Web of Science | Google Scholar . - Impact of persistent cortical activity (up states) on intracortical and thalamocortical synaptic inputs. J Neurophysiol 102: 119–131, 2009.
Link | Web of Science | Google Scholar . - Thalamocortical Up states: differential effects of intrinsic and extrinsic cortical inputs on persistent activity. J Neurosci 27: 4261–4272, 2007.
Crossref | PubMed | Web of Science | Google Scholar . - Sleep state switching. Neuron 68: 1023–1042, 2010.
Crossref | PubMed | Web of Science | Google Scholar . - Cell type-specific circuits of cortical layer IV spiny neurons. J Neurosci 23: 2961–2970, 2003.
Crossref | PubMed | Web of Science | Google Scholar . - Turning on and off recurrent balanced cortical activity. Nature 423: 288–293, 2003.
Crossref | PubMed | Web of Science | Google Scholar . - Anomalous rectification in neurons from cat sensorimotor cortex in vitro. J Neurophysiol 57: 1555–1576, 1987.
Link | Web of Science | Google Scholar . - Negative slope conductance due to a persistent subthreshold sodium current in cat neocortical neurons in vitro. Brain Res 236: 221–226, 1982.
Crossref | PubMed | Web of Science | Google Scholar . - Cholinergic and noradrenergic modulation of the slow (approximately 0.3 Hz) oscillation in neocortical cells. J Neurophysiol 70: 1385–1400, 1993a.
Link | Web of Science | Google Scholar . - Neuronal activities in brain-stem cholinergic nuclei related to tonic activation processes in thalamocortical systems. J Neurosci 10: 2541–2559, 1990.
Crossref | PubMed | Web of Science | Google Scholar . - A novel slow (<1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components. J Neurosci 13: 3252–3265, 1993b.
Crossref | PubMed | Web of Science | Google Scholar . - Stability of thalamocortical synaptic transmission across awake brain states. J Neurosci 29: 6851–6859, 2009.
Crossref | PubMed | Web of Science | Google Scholar . - Potassium currents in hippocampal pyramidal cells. Prog Brain Res 83: 161–187, 1990.
Crossref | PubMed | Web of Science | Google Scholar . - Barrel cortex microcircuits: thalamocortical feedforward inhibition in spiny stellate cells is mediated by a small number of fast-spiking interneurons. J Neurosci 26: 1219–1230, 2006.
Crossref | PubMed | Web of Science | Google Scholar . - Cerebral activity and behavior: control by central cholinergic and serotonergic systems. Intl Rev Neurobiol 30: 225–340, 1988.
Crossref | PubMed | Web of Science | Google Scholar . - Background synaptic activity is sparse in neocortex. J Neurosci 26: 8267–8277, 2006.
Crossref | PubMed | Web of Science | Google Scholar . - Cholinergic switching within neocortical inhibitory networks. Science 281: 985–988, 1998.
Crossref | PubMed | Web of Science | Google Scholar .