There are several stages of sleep that cycle repeatedly through the night with each producing distinctive patterns of electrical activity in the brain. It is thought that these patterns may help us to remember things that have happened throughout the day. Cells in parts of the brain called the hypothalamus and the brainstem control transitions between sleep stages. They regulate the release of chemicals known as neuromodulators in many parts of the brain, including the cortex and thalamus, which play the roles in memory and learning. Researchers now know how the neuromodulators influence the properties of individual brain cells. However, it is not clear how coordinated action of many neuromodulators result in the patterns of electrical activity seen in the brain during each stage of sleep.

Krishnan et al. used a computer model to investigate how three of these neuromodulators – acetylcholine, histamine and GABA – shift electrical activity in the brain between sleep stages. The computer model was able to recreate the network of brain cells in the cortex and thalamus and how this network responds to the changes in the levels of neuromodulators. The study found that simultaneous and balanced changes of acetylcholine, histamine, and GABA work together to shift the brain between the stages of sleep and to initiate patterns of the brain electrical activity specific to the different sleep stages.

Krishnan et al. predict that the relative differences in the level of acetylcholine in the brains of humans, cats and mice may explain why different species have different patterns of electrical activity during sleep. The study also found that an anesthetic drug called propofol may induce sleep-like patterns of electrical activity in the human brain by affecting the levels of all three of the neuromodulators. More studies are needed to look at how the networks of cells in the cortex and thalamus communicate with the brainstem, and how changes in the levels of neuromodulators affect memory and learning.

During sleep, unique electrophysiological rhythms are observed in the EEG, intracellular recordings, EOG and EMG (Steriade et al., 1993a; Niedermeyer and da Silva, 2005). These features form the basis for classification of sleep into different stages: rapid eye movement sleep (REM), and stages N1 (Stage 1), N2 (Stage 2) and N3 (Stage 3) of non-REM sleep (Rechtschaffen and Kales, 1968; Iber and Medicine, 2007; Silber et al., 2007). A typical night of sleep consists of 4–5 sleep cycles of transitions across stages in healthy adults. Each sleep cycle shows progression of sleep stages in the following order, N1 $\to$ N2 $\to$ N3 $\to$ N2 $\to$ REM (Feinberg and Floyd, 1979). Approximately 75% of sleep is spent in NREM stages. There is also a greater amount of N3 (also termed Slow Wave Sleep, SWS) earlier in the night, whereas REM stage is relatively dominant during later cycles of sleep (Aeschbach and Borbely, 1993).

Each sleep stage is characterized by specific patterns of the brain electrical activity. The N1 stage consists of slow eye movements and low-amplitude low-frequency (4–7 Hz) EEG activity in humans. The sleep stage N2 is dominated by sleep spindle oscillations at 10–17 Hz (in humans) with waxing-and-waning field potentials. Spindles last about 0.5–2 s and recur every 2–20 s. Similar events have been studied in non-human mammals, where their frequency may be as low as 7 Hz. During the N3 delta band (0.2–4 Hz), rhythms are found in the EEG (Rechtschaffen and Kales, 1968; Amzica and Steriade, 1998). This stage is dominated by the slow oscillation, consisting of active (Up) and silent (Down) cortical states alternating at frequency 0.2–1 Hz, that are prominent and visible in the EEG, and in the extracellular and intracellular recordings (Steriade et al., 1991, 1993a; 2001, Werth et al., 1996; Timofeev et al., 2000, 2001). During Up states, most cells within the cerebral cortex are relatively depolarized and may generate action potentials; during Down states, most cortical neurons are hyperpolarized and do not fire (Steriade et al., 1993b; Contreras and Steriade, 1995; Timofeev et al., 2000). The slow oscillation may nest faster spindles, which are commonly found during Down-to-Up state transitions (Molle et al., 2002; Clemens et al., 2007). During REM sleep, electrical brain activity resembles that of the awake state, including mixed signals with bursts of alpha (7–12 Hz) activity (Cantero and Atienza, 2000). Precise coordination of different EEG rhythms during sleep is believed to be critical for memory consolidation which manifests in reactivation of specific neural activity patterns – sleep replay – that have been observed in different brain areas including the hippocampus, amygdala, neocortex and striatum (Nadasdy et al., 1999; Pennartz et al., 2004; Euston et al., 2007; Popa et al., 2010; Bendor and Wilson, 2012).

Neuromodulators, such as acetylcholine (ACh), GABA, histamine (HA), serotonin (5-HT) and norepinephrine (NE), are known to vary significantly during sleep and awake as well as across sleep stages. Compared to the awake state, in N2 and N3, the levels of ACh and monoamines such as HA and NE are reduced while the level of the inhibitory neurotransmitter GABA is increased (Aston-Jones and Bloom, 1981); (McCormick, 1992; Vazquez and Baghdoyan, 2001; Lena et al., 2005; Vanini et al., 2011). During REM sleep, the level of ACh is increased, but monoamines and GABA are reduced (Baghdoyan and Lydic, 2012). While intracellular and synaptic targets of specific neuromodulators are somewhat known, we still lack a clear understanding of how the orchestrated action of many neuromodulators leads to the very specific types of the brain electrical activity in awake and sleep.

In this study, we present a comprehensive computational model of the thalamocortical system implementing effects of neuromodulators and identify the critical intrinsic and synaptic neuronal mechanisms required to explain transitions between sleep stages within an ultradian cycle. We predict that the differences in the temporal dynamics of spindle and delta band oscillations observed in the LFP recordings of cats and mice and ECoG activity of human subjects during SWS (N3 in humans) could arise from the relative differences in the level of neuromodulators. We further apply the model to explain electrical activity observed in propofol-induced anesthesia. Our study predicts that even minor changes of neuromodulators may affect the properties of sleep spindles and sleep slow oscillation in NREM sleep, thus possibly affecting the dynamics of memory consolidation.

Levels of the ACh, HA and GABA vary across sleep stages in vivo. During NREM stages (N2 and N3), the ACh and HA levels are reduced compared to awake; during REM, the ACh and HA levels are increased (Baghdoyan and Lydic, 2012) (Figure 1B). The level of the inhibitory neurotransmitter, GABA, is increased during N2 and N3 compared to awake, while reduced during REM sleep (McCormick, 1992; Vanini et al., 2011). To link the action of these neuromodulators to characteristic brain electrical activity in different sleep stages, we developed a thalamocortical network model implementing the basic effects of the ACh, HA and GABA (see 'Materials and methods'). The model included populations of pyramidal (PY) cells and inhibitory interneurons (IN) in the cortex, and thalamic relay (TC) and reticular nucleus (RE) neurons in the thalamus (Figure 1A); it was designed based on our previous studies of sleep slow oscillations and sleep spindles (Bazhenov et al., 1998, 1999; Timofeev et al., 2000; Bonjean et al., 2011; Chen et al., 2012; Wei et al., 2016).

Figure. 1

Download asset Open asset

Specific effects of neuromodulators were implemented by changing the strength of the intrinsic and synaptic currents in the cortical and thalamic neurons (see 'Materials and methods' for more details). While our model does not capture the entire spectrum of changes associated with known effects of neuromodulators, we identified a minimal set sufficient to account for characteristic changes of brain electrical activity across the sleep-wake cycle. Briefly, a reduction of ACh was implemented as an increase in the conductance of the ${\text{K}}^{+}$ leak current in TC, PY and IN neurons and a reduction of the ${\text{K}}^{+}$ leak current in RE cells (McCormick, 1992) (Figure 1C,D). The level of ACh also determined the maximal conductance of the excitatory AMPA connections (Figure 1C,D), so that reduction of ACh led to an increase in AMPA connection strength, in agreement with experimental findings (Gil et al., 1997; Kimura et al., 1999; Hsieh et al., 2000). The effect of HA was implemented as a shift in the activation curve of a hyperpolarization-activated cation current, ${\text{I}}_{\text{h}}$, in TC cells (Figure 1B) (McCormick and Williamson, 1991; McCormick, 1992a). Higher values of HA led to a positive shift in the ${\text{I}}_{\text{h}}$ activation curve (Figure 1B) as seen in vitro (McCormick and Williamson, 1991). The level of HA also determined the strength of the ${\text{K}}^{+}$ leak conductance in all neurons (see 'Materials and methods' section). The level of GABA determined the maximal conductance of the GABAergic synaptic connections. We also tested in the model the effect of extracellular GABA concentration on tonic inhibition.

Below we will first show that combined action of all these factors is sufficient to explain transitions between the awake state, the N2 and N3 stages and the REM sleep state. Next, we map the relationship between the space of electrical brain activities during sleep and the neurochemical space. We then discuss the impact of specific neuromodulators on brain electrical activity and sleep patterns. Next, we apply our model to explain discrepancies in the animal and human data regarding phase coupling between specific sleep rhythms. Finally, we apply our model to simulate the effect of propofol anesthesia on brain electrical activity.

Collective action of acetylcholine, histamine and GABA explains transitions between sleep and awake states

When the model implemented the intracellular and synaptic changes associated with the putative actions of ACh, HA and GABA, it produced the main features of the different sleep stages observed in human and animal studies. During the awake period, electrical activity in the cortical PY and IN and thalamic RE cells was sparse, that is mean firing was around 2–3 Hz and 4–5 Hz in PY and IN cells, respectively, while the TC neurons primarily remained silent and RE cells were busting intermittently (Figure 2A and Figure 3A). To test the stability of network dynamics, we applied brief thalamic stimulation in the form of a short (100 ms) DC pulse during the awake state. It triggered a transient increase of firing in cortical PY neurons and INs, after which the network returned to the baseline, as observed in experiments with a sensory input during alert awake conditions (Bruno and Sakmann, 2006; Hengen et al., 2016). The average activity (simulated LFP) during awake period (Figure 2C) showed no oscillatory activity or large deflections from baseline, reflecting desynchronized neuronal firing.

Figure 2

Download asset Open asset

Figure 3

Download asset Open asset

The N2 stage was characterized in the model by the periods of spindles – thalamically-organized bursts of 7–15 Hz oscillations lasting 0.5 to 2 s each and recurring every 3–20 s (Steriade et al., 1993a). In the model, we observed a transition to N2-like activity following reduction of ACh and HA, and increase of GABA. No external stimulation was applied to initiate or maintain spindles; however, thalamic stimulation could trigger a spindle response. In this state of the model (see N2 in Figure 2), periods of 7–15 Hz oscillatory activity reappeared spontaneously and lasted for at least 2–5 s as revealed by the spectrogram (Figure 2D). Membrane voltages were hyperpolarized in the PY, IN and TC cells due to the reduction of the ACh (Figures 2B and 3B), similar to the intracellular recording (Steriade et al., 1993c). In agreement with the prior intracellular data (Contreras and Steriade, 1996) and computational models (Bazhenov et al., 2000; Bonjean et al., 2011), spindles recurred every 3–10 s.

When the levels of ACh and HA were reduced and GABA was increased compared to the simulations of the N2 state, N3-like activity appeared in the network (Figures 2 and 3C). In this state, a stereotypical slow oscillation consisting of transitions between Up and Down cortical states at a frequency of around 0.5–1 Hz was observed in the LFP, similar to experimental data (Steriade et al., 1993c). Up states were characterized by the higher spiking activity in all cortical pyramidal cells, interneurons and thalamic reticular neurons, while TC cells showed a depolarized state (similar to intracellular recordings from higher order thalamic nuclei [Sheroziya and Timofeev, 2014]) with a few spikes usually at transition from the silent to the active state; the Down states corresponded to the periods of the network silence (Timofeev et al., 2000; Bazhenov et al., 2002; Compte et al., 2003). As in vivo, faster spindle oscillations were found during Up states of the slow oscillation in the model.

To model REM sleep, the level of ACh was increased slightly compared to the awake period, whereas the levels of HA and GABA were reduced. In this network state, there was an overall increase in the cortical activity with very brief periods of 7–15 Hz oscillations that appeared on the background of desynchronized low-frequency cortical firing. The clusters of synchronized firing were somewhat similar to that observed during spindle oscillations. However, they appeared to be more localized in space. In the model, these brief periods of 7–15 Hz synchronized firing (Figure 3D) depended on the reduced level of monoamines in the thalamus and higher excitability in the cortex. We propose that these periods of synchronized oscillations correspond to the alpha or mu rhythms (7–13 Hz activity) that have been reported during REM sleep (Cantero and Atienza, 2000).

Effect of neuromodulators on spindle and slow oscillation power and network synchronization

In order to reveal the specific role of different neuromodulators on network dynamics, the levels of ACh, GABA and HA were varied across a wide range (ACh: 0 to 120%, GABA: 25–225%, HA: 34 to 100% of the awake stage). Power was measured in the spindle (7–15 Hz) and delta (0.5–4 Hz) frequency bands, during 10 s periods for each combination of the neuromodulators. In our model, during N3 the delta band activity was dominated by slow oscillations (0.5–1 Hz). Synchronization among network sites was measured using the phase locking value (PLV) in a broadband frequency (0.5–20 Hz) range, between the LFPs obtained by averaging membrane voltages within five groups of neurons, each comprising 100 PY neurons. Figure 4A shows the summary of the results from all simulations projected to the spindle vs. delta power plane. Several clusters became apparent. Using cluster analysis (Gaussian mixture model) with spindle and delta power and PLV as inputs, 10 different clusters were identified (Akaike information criteria saturated around 10 components). The mean values of the clusters are indicated by the different color spheres in Figure 4A and examples of characteristic activity in each cluster are shown in Figure 4C.

Figure 4

Download asset Open asset

We found that identified cluster states corresponded to either well-defined characteristic types of sleep activity (e.g. sleep spindles or slow oscillations) or various mixed states. When spindle and delta power were both low (e.g. clusters 3 and 4), the activity resembled an awake or REM sleep. In many instances of activity in cluster 4 (Figure 4C), we observed very brief and spatially localized periods of 7–15 Hz oscillations, which were similar to the alpha or mu rhythms recorded during REM sleep. For higher levels of delta power, two groups of clusters were observed: clusters corresponding to low spindle power (e.g. cluster 2) and clusters where spindle power was positively correlated to the delta power (e.g. clusters 5, 6). Cluster 2 with low-spindle power but high-delta power corresponded to the stereotypic transitions between Up and Down states (Figure 4C) as observed during the sleep slow oscillation in vivo, while clusters with a correlated spindle and delta power (clusters 5 and 6) corresponded to the mixed states, where brief (cluster 6) or long (cluster 5) periods of spindles appeared between periods of slow oscillation. Finally, a distinct cluster 1 corresponded to low-delta power, but high-spindle power. This cluster represented quasi-continuous spindle activity (Figure 4C).

The clusters shown in the Figure 4A were identified based on the quantitative characteristics of the electrical activity in the network. In order to link them to specific levels of the neuromodulators, all detected clusters were projected to the neurochemical space. Such projection was performed by fitting an ellipsoid around the means of each cluster in the three-dimensional space of the neuromodulators (Figure 4B). Thus, location of each ellipsoid provides a guide to the range of neuromodulators that produced a specific electrographic pattern. A multivariate ANOVA on the location of the different clusters in the neuromodulator space was significant (p<10e-6), suggesting that the clusters occupy different regions. Clusters 3 and 4 (green ellipsoids in Figure 4B corresponding to the awake or REM sleep like electrical activity) were associated with relatively high ACh levels such as that observed during awake and REM periods. Cluster 2 (violet ellipsoid in Figure 4B corresponding to the slow oscillation) was observed in the low ACh conditions, as found during N3. Clusters 5 and 6 (sleep stage 2 based on electrical activity) corresponded to intermediate levels of the neuromodulators. Finally, cluster 1 was observed for lower HA and higher GABA levels, in agreement with the data showing that strong GABA promotes spindle activity (Parrino and Terzano, 1996). In sum, our analysis confirmed a specific role of neuromodulators in controlling the electrical activity of the thalamocortical network in different sleep and awake conditions, in agreement with past experimental studies (Baghdoyan and Lydic, 2012; McCormick, 1992).

Minimal models of sleep stage transitions

In order to further identify the critical mechanisms responsible for state transitions in the thalamocortical network, we tested the minimal changes of the neuromodulators that were sufficient to generate the essential electrographic characteristics of each sleep stage (Figure 5). Changes in the neuromodulators were restricted to either thalamus or cortex to isolate the role of different regions. Furthermore, we focused on manipulations of the ACh and HA and not GABA, since in our model GABA modulation (within the physiological range) alone was not able to induce any state transitions. We found that in order to induce a transition from awake to N2, a reduction of HA in the thalamus was necessary but produced very focal and scattered spindles (Figure 5C). Reduction of both ACh and HA in the thalamus (Figure 5D) resulted in spindle activity that was similar to that in a full model (Figure 5A, see also Figure 3B). Cortical changes alone did not lead to spindles (Figure 5B) but were required to obtain cortical membrane potentials in the physiological range during N2. We concluded that reduction of ACh and HA in the thalamus, along with reduction of ACh in the cortex, constitute a minimal model needed to simulate the main features of the transition from the awake state to the N2 sleep state.

Figure 5

Download asset Open asset

Transition from N2 to N3 could be observed when ACh was reduced in the cortex alone (Figure 5F). In contrast, even when all the other proposed neuromodulatory changes (reduction of ACh and HA, and increase of GABA) were applied in the thalamus, the thalamocortical network did not display N3-like activity (Figure 5E), suggesting that appearance of the slow oscillation during N3 requires neuromodulatory action within the cortex itself. Since reduction of the ACh has two main actions in our model, an increase in excitatory connection strength and an increase of leak currents, we examined if either of these changes alone could result in transition to the N3. Increasing excitatory AMPA conductance alone led to an activated state (Figure 5G), and an increase in the ${\text{K}}^{+}$ leak conductance alone led to a hyperpolarized quiescent state in the network (Figure 5H). Thus, both actions together were required for transition between the N2 and N3.

Comparison between sleep stages in cats, mice and humans

We next examined if our thalamocortical model implementing varying levels of neuromodulators can explain LFP data from regions of cat and mouse neocortex across sleep stages (Figure 6 A,B). Surprisingly, visual scoring of cat LFPs revealed that there were only relatively short periods of the S2 (equivalent to N2; to describe animal experiments we use here terminology common in the animal studies) compared to the SWS (equivalent to N3). This result was consistent across animals. The power in the spindle (8–15 Hz) and delta (0.2–4 Hz) frequency bands was measured across all channels and the average power was plotted in Figure 6C and D, separately for the cortical and thalamic (VPL) recordings. The amplitude of the slow oscillation and spindles were both high during the S2 and SWS, and were reduced in awake and REM. This suggests that the SWS in naturally sleeping cats is marked with the relatively high-spindle activity. Slow (0.2–1 Hz) oscillations provided the major contribution to delta band activity during SWS. To further examine the interaction between sleep spindles and slow oscillations, we plotted the power in the spindle vs delta frequency bands across different cortical LFP electrodes (Figure 6E). We found a strong correlation between slow oscillation and spindle activity. In mice (Figure 6F,G), a similar relationship between spindle and delta power was observed across recordings from five animals (3 channels per mouse); a positive correlation was observed in majority of the channels (9 channels out of 15 channels showed significant correlation of which 7 were positively correlated [r ranged from 0.17 to 0.43; p<0.05] and 2 were negatively correlated [r was −0.41 and –0.43; p<0.05]). Recordings from four out of five mice revealed significant correlations. In one mouse (Figure 6F–G), frontal and posterior somatosensory electrodes showed a significant positive correlation between slow oscillations and spindle activity.

Figure 6

Download asset Open asset

Our analysis of the neuromodulator space (Figure 4) suggests that a moderate reduction of ACh would lead to a network state with correlated variations in spindle and delta power (clusters 4, 5 and 6 in Figure 4A). Thus, in order to replicate our animal data (cat and mouse LFP activity), a computational model with a relatively high level of ACh during SWS (reduction to 75% of the baseline awake level) was required (Figure 7A). In contrast, in our control simulations of SWS (N3), the ACh was reduced to 57% of the baseline (Figures 2 and 3). Using the new model with the higher level of the ACh during SWS, we found a relative increase in the spindle activity during the SWS (Figure 7A) and a strong positive correlation between slow oscillations and spindle activity (Figure 7B, right), as was observed in the animal LFP data (Figure 7B, left).

Figure 7

Download asset Open asset

Next, ECoG recordings from humans during different sleep stages were examined for the relationship between delta (0.01–2 Hz) and spindle (9–17 Hz) power (Figure 8A,B). (Note, these slightly different frequency bands are conventional in human studies.) In contrast to the pattern found in cats and mice, there was a significant negative correlation between the spindle and the delta power observed in successive 30 s epochs obtained during the N2 and N3 (Figure 8C and D) in humans. Generally, the slow oscillation was higher during N3 than during N2, whereas spindle activity was higher in N2 than in N3, resulting in a negative correlation across epochs. The average Pearson’s correlation coefficient across 42 channels was −0.2595 (sd = 0.2281) (Figure 8E). Fifteen out of 42 electrodes showed a significant negative correlation (p<0.001), and no electrodes showed a significant positive correlation. We also observed periods of increased power in the 7–15 Hz frequency band during REM and awake suggesting transient periods of alpha (or mu) rhythm. To simulate the negative correlation between sleep slow oscillation and spindle activities, we again applied results from our analysis of the neuromodulator space (Figure 4). When the ACh level was reduced to 45% of the baseline during the N3 (compared to 57% used in the control model, see Figures 2, 3), the network activity revealed a negative correlation between the delta and spindle frequency bands (Figure 8 F and G). These findings suggest that a relatively low level of the ACh during the N3 could possibly explain the negative correlation between the slow oscillation and spindle activities as observed in human ECoG recordings.

Figure 8

Download asset Open asset

Modeling the effects of propofol anesthesia requires the combined action of ACh, HA and GABA

We next applied our model to explain the features of the electrical activity induced by propofol anesthesia. In humans, propofol has shown to increase 8–15 Hz oscillations frontally and to reduce alpha oscillations posteriorly (Murphy et al., 2011; Purdon et al., 2013). Propofol application also increased the slow oscillations in the delta (1–4 Hz) frequency range across all regions. The onset of the loss of consciousness was closely matched by the increase of slow (<1 Hz) and 8–15 Hz oscillations (Purdon et al., 2013). We examined if our model could capture these changes in the EEG due to the action of propofol.

Propofol is a GABA agonist and is known to increase the decay time constant of the GABA-A mediated inhibitory post-synaptic potential (Kitamura et al., 2003). Multiple lines of evidence suggest that propofol also acts by reducing the action of ACh and HA. Indeed, propofol is shown to decrease the level of ACh in frontal cortex (Kikuchi et al., 1998) and to attenuate ACh receptor responses (Flood et al., 1997; Murasaki et al., 2003). Increasing ACh transmission prevents the action of propofol in humans (Meuret et al., 2000). Propofol inhibits the activity of the tuberomammillary nucleus (Nelson et al., 2002), although its action on ventrolateral preoptic nucleus neurons (Liu et al., 2013) leading to reduction of HA.

When these effects of ACh and HA were implemented in our model, as described before, and the decay time constant of GABA was increased, we observed a large increase in slow oscillations as well as oscillatory activity in 8–15 Hz frequency band (Figure 9A). In comparison to the model of natural sleep (Figure 9C), there was elevated 8–15 Hz power in the propofol condition (Figure 9D). This is consistent with our recordings from cats (not shown), where we found that 8–15 Hz power under propofol was high compared to natural sleep or ketamine-xylazine anesthesia. These results are also consistent with the previous computational models of propofol application, where the increase in decay time constant of IPSPs led to increase in 8–15 Hz oscillations (Ching et al., 2010). However, in our model, when only the time constant of the GABA-A IPSP was increased compare to the awake condition, the model failed to generate slow oscillations or the 8–15 Hz synchronized activity (Figure 9B). Overall, our study suggests that the known action of propofol in vivo may require its indirect effect on both the ACh and HA neuromodulatory systems.

Figure 9

Download asset Open asset

Role of GABA in sleep stage transitions

Experimental studies suggest that the concentration of extracellular GABA in the neocortex is higher during NREM sleep and it is reduced during REM sleep compared to the awake state (Vanini et al., 2012). Increase of extracellular GABA may reflect an increase of synaptic inhibition. Indeed, during slow wave sleep, synaptic excitation and inhibition are balanced (Rudolph et al., 2007; Haider et al., 2006). Since decrease of ACh during stage N2/N3 sleep is known to increase excitatory connections (Gil et al., 1997), synaptic inhibition and thus phasic GABA release may be also increased, which would then be reflected in elevated extracellular GABA. The effect of increase of synaptic inhibition was implemented in our baseline model as reported above.

Nevertheless, the origin of the change in the extracellular GABA is still not fully understood. Increase of extracellular GABA may increase tonic inhibition but not necessarily be associated with increase of the phasic GABA release. Further, several studies revealed co-release of GABA with ACh and HA (Saunders et al., 2015; Yu et al., 2015) which suggests a decrease of GABA during NREM sleep while the observations from microdialysis experiments (Vanini et al., 2012) suggest an increase of GABA during NREM sleep. Therefore, in our study, we also tested models with (a) no change in the level of the GABA release; (b) increase of tonic inhibition reflecting increase of the extracellular GABA, based on the observations from microdialysis experiments (Vanini et al., 2012); and (c) tonic inhibition proportional to the ACh and HA levels, based on the evidence of the co-release of GABA with ACh (Saunders et al., 2015) and HA (Yu et al., 2015). In all these conditions (Figure 1), we found transitions between sleep stages similar to those reported in the baseline model. However, the model implementing co-release of GABA with ACh had poorly formed spindle activity during N2 and elevated alpha activity during REM sleep. Future experiments are required to distinguish between changes in tonic and phasic inhibition to match specific features of the synchronized oscillatory activity across sleep stages.

Figure 10

Download asset Open asset

The functional state and the patterns of electrical activity of human and animal brain are continuously influenced by a broad range of neuromodulators (McCormick, 1992; Steriade et al., 2001; Baghdoyan and Lydic, 2012). Changes in the level of neuromodulators have been correlated with sleep induction as well as transitions between characteristic EEG sleep patterns (McCormick, 1992; McCormick and Bal, 1997; Steriade et al., 2001). However, the mechanisms whereby specific cellular and synaptic changes triggered by the combined action of neuromodulators result in transitions between sleep stages and their precise electrical characteristics remain poorly understood. In this new study, we used biophysically realistic computational models of the thalamocortical system to identify the minimal set of cellular and network level changes, linked to the putative action of the neuromodulators, that is sufficient to explain the characteristic neuronal dynamics during sleep as well as the transitions between primary sleep stages and from sleep to awake state. Our study predicts a critical role of the neuromodulators in controlling the precise spatio-temporal characteristics of neuronal synchronization that manifest in major sleep rhythms – sleep spindles and slow oscillations. Our results are consistent with in vivo intracellular and LFP recordings from animals and ECoG data from human subjects, as reported in this study.

Previous studies identified intrinsic and synaptic mechanisms involved in generating specific types of sleep rhythms. The survival of the slow oscillation (<1 Hz global electrical activity found during deep (N3) sleep) after extensive thalamic lesions in vivo (Steriade et al., 1993b), its existence in cortical slice preparations (Sanchez-Vives and McCormick, 2000), and the absence of the slow oscillations in the thalamus of decorticated cats (Timofeev and Steriade, 1996) point to the primarily intracortical origin for this rhythm. (Note that thalamus can be actively involved in synchronization of the SO [Lemieux et al., 2014]). In vivo and in vitro studies suggest that the minimal substrate accounting for spindle oscillations (7–15 Hz recurrent oscillatory activity found during stage 2 (N2) of NREM sleep) is the interaction between thalamic reticular and relay cells (Steriade and Deschénes, 1984; Steriade et al., 1985; Steriade and Llinas, 1988, 1990; von Krosigk et al., 1993). Previous computational studies proposed a minimal set of the mechanisms sufficient to model sleep spindles and slow oscillations (Destexhe et al., 1996; Bazhenov et al., 1998, 1999; Timofeev et al., 2000; Bazhenov et al., 2002; Compte et al., 2003; Hill and Tononi, 2005; Bonjean et al., 2011). An increase in the K⁺ leak current was identified as a critical component for the transition between awake and slow wave sleep (Bazhenov et al., 2002; Hill and Tononi, 2005), and it was predicted that synchronization of the slow oscillation depends on cortico-cortical connections (Hill and Tononi, 2005). Several phenomenological and reduced mathematical models have been proposed to test the effects of neuromodulators on the sleep-wake cycle (McCarley and Hobson, 1975; Robinson et al., 2011; Schellenberger Costa et al., 2016). Nevertheless, none of these previous models examined transitions among all major sleep stages based on biophysical mechanisms.

In this study, we found that reduction in HA and ACh in the thalamus was required for induction of N2-like activity, characterized by recurrent spindles with characteristics consistent with animal and human data. Other neuromodulatory changes (such as change of GABA conductance) were not sufficient alone to induce N2-like activity, but played a role in determining the power and the synchrony of oscillations. Reduction in the ACh levels within the cortical network was identified to be a minimal requirement for the transition from the N2 (spindles) to the N3 (slow oscillation). Our study of the minimal models revealed the importance of the cell-type-specific action of the neuromodulators. Indeed, in order to model sleep stage transitions, the crucial changes were in HA, primarily acting on thalamic neurons, and in ACh, primarily acting within a cortex. All the network rhythms reported in this study resulted from intrinsic (autonomous) dynamics of the thalamocortical network; no external stimulation was applied to entrain any of the oscillatory activities in the model.

By parametrically varying the levels of the ACh, HA and GABA, we observed a wider range of activities in the thalamocortical network than have been previously reported in the computational models, but which do correspond to those that have been observed experimentally. While we found network states dominated by either spindles or slow oscillations, under certain conditions (such as only moderate reduction of the ACh level) we also observed mixed states combining these two major rhythms. Such mixed states may occur in vivo under normal conditions (Aeschbach and Borbely, 1993; Muller et al., 2006); our recent behavior study suggests that the strengths of the phase amplitude coupling between the spindle and the slow oscillation during NREM sleep correlates with memory consolidation (Niknazar et al., 2015). We speculate that the mixed states may become prevalent in some pathological conditions such as in Alzheimer’s disease where sleep is altered and changes to the neuromodulatory system are reported (Wulff et al., 2010). Furthermore, our results on the putative relationship between levels of neuromodulators and characteristic sleep electrical activities could be potentially relevant to understanding the changes observed with aging. Indeed, the level of the HA in cerebrospinal fluid is known to be elevated with aging (Prell et al., 1988), where spindle power is known to be reduced (Crowley et al., 2002). In our model, the HA level is one of the major predictors of spindle power. Thus, we can speculate that an increase of HA with age could contribute to the decline of the spindle power.

In our study, the effects of neuromodulators were limited by changing the following four model properties – maximal conductances of the K⁺-leak, I_h, GABA and AMPA currents. We found that specific combinations of changes of these core parameters were sufficient to cause transitions between sleep stages. We also tested a more detailed model, implementing the effects of ACh and HA on the other K⁺ currents as well as on the persistent Na⁺ and potassium M-currents (Constanti and Galvan, 1983; McCormick and Williamson, 1989; Mittmann and Alzheimer, 1998). We observed that this extended model qualitatively captured results similar to the minimal model. In contrast, after fixing the K⁺-leak, I_h, GABA and AMPA currents but leaving the action of neuromodulators on the other channels, the model failed to show transitions between sleep stages, suggesting that the cellular and synaptic properties identified in our study are critical in determining patterns of electrical activity in the thalamocortical network.

In developing our model, we assumed that the levels of ACh, GABA and HA were different between N2 (Stage 2) and N3 (Stage 3 or SWS). These assumptions are consistent with data on spiking activity in the Basal Forebrain region (Aston-Jones and Bloom, 1981; Szymusiak et al., 2000), suggesting a difference in ACh between the N2 and N3. Another assumption used in this study was that ACh modulates the strength of the excitatory AMPA synapses in the intracortical and thalamocortical circuits. Indeed, ACh suppresses the spread of excitation in vivo (Kimura et al., 1999) and reduces evoked responses (Chauvette et al., 2012). In slice experiments, ACh activation of muscarinic receptors suppressed both intracortical and thalamocortical excitatory connections, while nicotinic activation led to suppression of the intracortical connections and enhancement of the thalamocortical connections (Gil et al., 1997; Hsieh et al., 2000).

By comparing recordings from animals (cats and mice LFP) and human ECoG data across sleep stages, we found a significant difference in the interaction between major sleep rhythms (spindles and slow oscillation) during NREM sleep. In cats and mice, there was a strong positive correlation between spindle and delta power, while in humans there was a negative correlation between oscillations in these frequency bands. This difference may reflect the prominence of N2 (stage 2) in humans, in whom N2 has the longest duration of any sleep stage. In contrast, in cats, N2 is very rare compared to N3. In humans, N2 is mainly characterized by spindles with relatively short periods of slow oscillation or isolated K-complexes. (K-Complexes are large downward deflections in EEG often followed by spindle activity and is one of the electrophysiological markers of N2 (Loomis et al., 1937; Amzica and Steriade, 1997; Cash et al., 2009). In N3, spindles continue to occur in conjunction with the slow oscillation, but the EEG is dominated by slow activity. Our model predicts that this inter-species difference can be explained by a difference in the relative level of ACh during NREM sleep. This result is consistent with microdialysis measurements in cats where ACh was present during SWS albeit at low concentrations compared to wakefulness (Marrosu et al., 1995).

Surprisingly, we found that even minor changes of neuromodulators may lead to significant changes in the characteristics of brain electrical activity during sleep. Our model predicts that fluctuations of the level of ACh over time may explain the appearance of short periods of slow oscillations in N2 in humans (Achermann and Borbely, 1997) or modulate gamma activity during Up-states of the slow oscillation (Mena-Segovia et al., 2008). We hypothesize that transient activities such as K-Complexes during N2 sleep or PGO waves during REM stage may arise from fluctuations in ACh and/or other neuromodulators. Ultimately, our study predicts that transient changes in the level of neuromodulators may affect the interaction and characteristic properties of sleep rhythms and may thus affect functional outcomes of sleep, such as the impact of sleep on memory and learning.

1. 1. Achermann P
   2. Borbély AA

   (1997) Low-frequency (< 1 Hz) oscillations in the human sleep electroencephalogram

   Neuroscience 81:213–222.

   https://doi.org/10.1016/S0306-4522(97)00186-3

   • PubMed
   • Google Scholar
2. 1. Aeschbach D
   2. Borbély AA

   (1993) All-night dynamics of the human sleep EEG

   Journal of Sleep Research 2:70–81.

   https://doi.org/10.1111/j.1365-2869.1993.tb00065.x

   • PubMed
   • Google Scholar
3. 1. Akaike H

   (1974) A new look at the statistical model identification

   IEEE Transactions on Automatic Control 19:716–723.

   https://doi.org/10.1109/TAC.1974.1100705

   • Google Scholar
4. 1. Amzica F
   2. Steriade M

   (1997) The K-complex: its slow (<1-Hz) rhythmicity and relation to delta waves

   Neurology 49:952–959.

   https://doi.org/10.1212/wnl.49.4.952

   • PubMed
   • Google Scholar
5. 1. Amzica F
   2. Steriade M

   (1998) Electrophysiological correlates of sleep delta waves

   Electroencephalography and Clinical Neurophysiology 107:69–83.

   https://doi.org/10.1016/S0013-4694(98)00051-0

   • PubMed
   • Google Scholar
6. 1. Aston-Jones G
   2. Bloom FE

   (1981)

   Activity of norepinephrine-containing locus coeruleus neurons in behaving rats anticipates fluctuations in the sleep-waking cycle

   Journal of Neuroscience 1:876–886.

   • PubMed
   • Google Scholar
7. 1. Baghdoyan HA
   2. Lydic R

   (2012)

   Basic Neurochemistry

   982–999, Chapter 57 - The neurochemistry of sleep and wakefulness, Basic Neurochemistry, 8th edn, New York, Academic Press.

   • Google Scholar
8. 1. Bazhenov M
   2. Timofeev I
   3. Steriade M
   4. Sejnowski T

   (2000)

   Spiking-bursting activity in the thalamic reticular nucleus initiates sequences of spindle oscillations in thalamic networks

   Journal of Neurophysiology 84:1076–1087.

   • PubMed
   • Google Scholar
9. 1. Bazhenov M
   2. Timofeev I
   3. Steriade M
   4. Sejnowski TJ

   (1998)

   Computational models of thalamocortical augmenting responses

   Journal of Neuroscience 18:6444–6465.

   • PubMed
   • Google Scholar
10. 1. Bazhenov M
    2. Timofeev I
    3. Steriade M
    4. Sejnowski TJ

    (1999) Self-sustained rhythmic activity in the thalamic reticular nucleus mediated by depolarizing GABAA receptor potentials

    Nature Neuroscience 2:168–174.

    https://doi.org/10.1038/5729

    • PubMed
    • Google Scholar
11. 1. Bazhenov M
    2. Timofeev I
    3. Steriade M
    4. Sejnowski TJ

    (2002)

    Model of thalamocortical slow-wave sleep oscillations and transitions to activated States

    Journal of Neuroscience 22:8691–8704.

    • PubMed
    • Google Scholar
12. 1. Bendor D
    2. Wilson MA

    (2012) Biasing the content of hippocampal replay during sleep

    Nature Neuroscience 15:1439–1444.

    https://doi.org/10.1038/nn.3203

    • PubMed
    • Google Scholar
13. 1. Bonjean M
    2. Baker T
    3. Lemieux M
    4. Timofeev I
    5. Sejnowski T
    6. Bazhenov M

    (2011) Corticothalamic feedback controls sleep spindle duration in vivo

    Journal of Neuroscience 31:9124–9134.

    https://doi.org/10.1523/JNEUROSCI.0077-11.2011

    • PubMed
    • Google Scholar
14. 1. Bruno RM
    2. Sakmann B

    (2006) Cortex is driven by weak but synchronously active thalamocortical synapses

    Science 312:1622–1627.

    https://doi.org/10.1126/science.1124593

    • PubMed
    • Google Scholar
15. 1. Cantero JL
    2. Atienza M

    (2000) Alpha burst activity during human REM sleep: descriptive study and functional hypotheses

    Clinical Neurophysiology 111:909–915.

    https://doi.org/10.1016/S1388-2457(99)00318-1

    • PubMed
    • Google Scholar
16. 1. Cash SS
    2. Halgren E
    3. Dehghani N
    4. Rossetti AO
    5. Thesen T
    6. Wang C
    7. Devinsky O
    8. Kuzniecky R
    9. Doyle W
    10. Madsen JR
    11. Bromfield E
    12. Eross L
    13. Halász P
    14. Karmos G
    15. Csercsa R
    16. Wittner L
    17. Ulbert I

    (2009) The human K-complex represents an isolated cortical down-state

    Science 324:1084–1087.

    https://doi.org/10.1126/science.1169626

    • PubMed
    • Google Scholar
17. 1. Chauvette S
    2. Seigneur J
    3. Timofeev I

    (2012) Sleep oscillations in the thalamocortical system induce long-term neuronal plasticity

    Neuron 75:1105–1113.

    https://doi.org/10.1016/j.neuron.2012.08.034

    • PubMed
    • Google Scholar
18. 1. Chen JY
    2. Chauvette S
    3. Skorheim S
    4. Timofeev I
    5. Bazhenov M

    (2012) Interneuron-mediated inhibition synchronizes neuronal activity during slow oscillation

    The Journal of Physiology 590:3987–4010.

    https://doi.org/10.1113/jphysiol.2012.227462

    • PubMed
    • Google Scholar
19. 1. Ching S
    2. Cimenser A
    3. Purdon PL
    4. Brown EN
    5. Kopell NJ

    (2010) Thalamocortical model for a propofol-induced -rhythm associated with loss of consciousness

    PNAS 107:22665–22670.

    https://doi.org/10.1073/pnas.1017069108

    • Google Scholar
20. 1. Clemens Z
    2. Mölle M
    3. Eross L
    4. Barsi P
    5. Halász P
    6. Born J

    (2007) Temporal coupling of parahippocampal ripples, sleep spindles and slow oscillations in humans

    Brain 130:2868–2878.

    https://doi.org/10.1093/brain/awm146

    • PubMed
    • Google Scholar
21. 1. Compte A
    2. Sanchez-Vives MV
    3. McCormick DA
    4. Wang X-J

    (2003) Cellular and network mechanisms of slow oscillatory activity (<1 Hz) and wave propagations in a cortical network model

    Journal of Neurophysiology 89:2707–2725.

    https://doi.org/10.1152/jn.00845.2002

    • Google Scholar
22. 1. Constanti A
    2. Galvan M

    (1983) M-current in voltage-clamped olfactory cortex neurones

    Neuroscience Letters 39:65–70.

    https://doi.org/10.1016/0304-3940(83)90166-0

    • PubMed
    • Google Scholar
23. 1. Contreras D
    2. Steriade M

    (1995)

    Cellular basis of EEG slow rhythms: a study of dynamic corticothalamic relationships

    Journal of Neuroscience 15:604–622.

    • PubMed
    • Google Scholar
24. 1. Contreras D
    2. Steriade M

    (1996) Spindle oscillation in cats: the role of corticothalamic feedback in a thalamically generated rhythm

    The Journal of Physiology 490 :159–179.

    https://doi.org/10.1113/jphysiol.1996.sp021133

    • PubMed
    • Google Scholar
25. 1. Crowley K
    2. Trinder J
    3. Kim Y
    4. Carrington M
    5. Colrain IM

    (2002) The effects of normal aging on sleep spindle and K-complex production

    Clinical Neurophysiology 113:1615–1622.

    https://doi.org/10.1016/S1388-2457(02)00237-7

    • PubMed
    • Google Scholar
26. 1. Destexhe A
    2. Bal T
    3. McCormick DA
    4. Sejnowski TJ

    (1996)

    Ionic mechanisms underlying synchronized oscillations and propagating waves in a model of ferret thalamic slices

    Journal of Neurophysiology 76:2049–2070.

    • PubMed
    • Google Scholar
27. 1. Destexhe A
    2. Mainen ZF
    3. Sejnowski TJ

    (1994) An efficient method for computing synaptic conductances based on a kinetic model of receptor binding

    Neural Computation 6:14–18.

    https://doi.org/10.1162/neco.1994.6.1.14

    • Google Scholar
28. 1. Euston DR
    2. Tatsuno M
    3. McNaughton BL

    (2007) Fast-forward playback of recent memory sequences in prefrontal cortex during sleep

    Science 318:1147–1150.

    https://doi.org/10.1126/science.1148979

    • PubMed
    • Google Scholar
29. 1. Feinberg I
    2. Floyd TC

    (1979) Systematic trends across the night in human sleep cycles

    Psychophysiology 16:283–291.

    https://doi.org/10.1111/j.1469-8986.1979.tb02991.x

    • PubMed
    • Google Scholar
30. 1. Flood P
    2. Ramirez-Latorre J
    3. Role L

    (1997) Alpha 4 beta 2 neuronal nicotinic acetylcholine receptors in the central nervous system are inhibited by isoflurane and propofol, but alpha 7-type nicotinic acetylcholine receptors are unaffected

    Anesthesiology 86:859–865.

    https://doi.org/10.1097/00000542-199704000-00016

    • PubMed
    • Google Scholar
31. 1. Gil Z
    2. Connors BW
    3. Amitai Y

    (1997) Differential regulation of neocortical synapses by neuromodulators and activity

    Neuron 19:679–686.

    https://doi.org/10.1016/S0896-6273(00)80380-3

    • PubMed
    • Google Scholar
32. 1. Haider B
    2. Duque A
    3. Hasenstaub AR
    4. McCormick DA

    (2006) Neocortical network activity in vivo is generated through a dynamic balance of excitation and inhibition

    Journal of Neuroscience 26:4535–4545.

    https://doi.org/10.1523/JNEUROSCI.5297-05.2006

    • PubMed
    • Google Scholar
33. 1. Hengen KB
    2. Torrado Pacheco A
    3. McGregor JN
    4. Van Hooser SD
    5. Turrigiano GG

    (2016) Neuronal fring rate homeostasis is inhibited by sleep and promoted by wake

    Cell 165:180–191.

    https://doi.org/10.1016/j.cell.2016.01.046

    • PubMed
    • Google Scholar
34. 1. Hill S
    2. Tononi G

    (2005) Modeling sleep and wakefulness in the thalamocortical system

    Journal of Neurophysiology 93:1671–1698.

    https://doi.org/10.1152/jn.00915.2004

    • PubMed
    • Google Scholar
35. 1. Hsieh CY
    2. Cruikshank SJ
    3. Metherate R

    (2000) Differential modulation of auditory thalamocortical and intracortical synaptic transmission by cholinergic agonist

    Brain Research 880:51–64.

    https://doi.org/10.1016/S0006-8993(00)02766-9

    • PubMed
    • Google Scholar
36. 1. Iber C
    2. AAoS M

    (2007)

    The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications: American Academy of Sleep Medicine

    The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications: American Academy of Sleep Medicine.

    • Google Scholar
37. 1. Kikuchi T
    2. Wang Y
    3. Sato K
    4. Okumura F

    (1998) In vivo effects of propofol on acetylcholine release from the frontal cortex, hippocampus and striatum studied by intracerebral microdialysis in freely moving rats

    British Journal of Anaesthesia 80:644–648.

    https://doi.org/10.1093/bja/80.5.644

    • PubMed
    • Google Scholar
38. 1. Kimura F
    2. Fukuda M
    3. Tsumoto T

    (1999) Acetylcholine suppresses the spread of excitation in the visual cortex revealed by optical recording: possible differential effect depending on the source of input

    European Journal of Neuroscience 11:3597–3609.

    https://doi.org/10.1046/j.1460-9568.1999.00779.x

    • PubMed
    • Google Scholar
39. 1. Kitamura A
    2. Marszalec W
    3. Yeh JZ
    4. Narahashi T

    (2003) Effects of halothane and propofol on excitatory and inhibitory synaptic transmission in rat cortical neurons

    Journal of Pharmacology and Experimental Therapeutics 304:162–171.

    https://doi.org/10.1124/jpet.102.043273

    • PubMed
    • Google Scholar
40. 1. Lemieux M
    2. Chen JY
    3. Lonjers P
    4. Bazhenov M
    5. Timofeev I

    (2014) The impact of cortical deafferentation on the neocortical slow oscillation

    Journal of Neuroscience 34:5689–5703.

    https://doi.org/10.1523/JNEUROSCI.1156-13.2014

    • PubMed
    • Google Scholar
41. 1. Liu YW
    2. Zuo W
    3. Ye JH

    (2013) Propofol stimulates noradrenalin-inhibited neurons in the ventrolateral preoptic nucleus by reducing GABAergic inhibition

    Anesthesia & Analgesia 117:358–363.

    https://doi.org/10.1213/ANE.0b013e318297366e

    • PubMed
    • Google Scholar
42. 1. Loomis AL
    2. Harvey EN
    3. Hobart GA

    (1937) Cerebral states during sleep, as studied by human brain potentials

    Journal of Experimental Psychology 21:127–144.

    https://doi.org/10.1037/h0057431

    • Google Scholar
43. 1. Léna I
    2. Parrot S
    3. Deschaux O
    4. Muffat-Joly S
    5. Sauvinet V
    6. Renaud B
    7. Suaud-Chagny MF
    8. Gottesmann C

    (2005) Variations in extracellular levels of dopamine, noradrenaline, glutamate, and aspartate across the sleep--wake cycle in the medial prefrontal cortex and nucleus accumbens of freely moving rats

    Journal of Neuroscience Research 81:891–899.

    https://doi.org/10.1002/jnr.20602

    • PubMed
    • Google Scholar
44. 1. Marrosu F
    2. Portas C
    3. Mascia MS
    4. Casu MA
    5. Fà M
    6. Giagheddu M
    7. Imperato A
    8. Gessa GL

    (1995) Microdialysis measurement of cortical and hippocampal acetylcholine release during sleep-wake cycle in freely moving cats

    Brain Research 671:329–332.

    https://doi.org/10.1016/0006-8993(94)01399-3

    • PubMed
    • Google Scholar
45. 1. McCarley RW
    2. Hobson JA

    (1975) Neuronal excitability modulation over the sleep cycle: a structural and mathematical model

    Science 189:58–60.

    https://doi.org/10.1126/science.1135627

    • PubMed
    • Google Scholar
46. 1. McCormick DA
    2. Bal T

    (1997) Sleep and arousal: thalamocortical mechanisms

    Annual Review of Neuroscience 20:185–215.

    https://doi.org/10.1146/annurev.neuro.20.1.185

    • PubMed
    • Google Scholar
47. 1. McCormick DA
    2. Williamson A

    (1989) Convergence and divergence of neurotransmitter action in human cerebral cortex

    PNAS 86:8098–8102.

    https://doi.org/10.1073/pnas.86.20.8098

    • PubMed
    • Google Scholar
48. 1. McCormick DA
    2. Williamson A

    (1991)

    Modulation of neuronal firing mode in cat and guinea pig LGNd by histamine: possible cellular mechanisms of histaminergic control of arousal

    Journal of Neuroscience 11:3188–3199.

    • PubMed
    • Google Scholar
49. 1. McCormick DA

    (1992) Neurotransmitter actions in the thalamus and cerebral cortex and their role in neuromodulation of thalamocortical activity

    Progress in Neurobiology 39:337–388.

    https://doi.org/10.1016/0301-0082(92)90012-4

    • PubMed
    • Google Scholar
50. 1. Mena-Segovia J
    2. Sims HM
    3. Magill PJ
    4. Bolam JP

    (2008) Cholinergic brainstem neurons modulate cortical gamma activity during slow oscillations

    The Journal of Physiology 586:2947–2960.

    https://doi.org/10.1113/jphysiol.2008.153874

    • PubMed
    • Google Scholar
51. 1. Meuret P
    2. Backman SB
    3. Bonhomme V
    4. Plourde G
    5. Fiset P

    (2000) Physostigmine reverses propofol-induced unconsciousness and attenuation of the auditory steady state response and bispectral index in human volunteers

    Anesthesiology 93:708–717.

    https://doi.org/10.1097/00000542-200009000-00020

    • PubMed
    • Google Scholar
52. 1. Mittmann T
    2. Alzheimer C

    (1998)

    Muscarinic inhibition of persistent Na+ current in rat neocortical pyramidal neurons

    Journal of Neurophysiology 79:1579–1582.

    • PubMed
    • Google Scholar
53. 1. Murasaki O
    2. Kaibara M
    3. Nagase Y
    4. Mitarai S
    5. Doi Y
    6. Sumikawa K
    7. Taniyama K

    (2003) Site of action of the general anesthetic propofol in muscarinic M1 receptor-mediated signal transduction

    Journal of Pharmacology and Experimental Therapeutics 307:995–1000.

    https://doi.org/10.1124/jpet.103.055772

    • PubMed
    • Google Scholar
54. 1. Murphy M
    2. Bruno MA
    3. Riedner BA
    4. Boveroux P
    5. Noirhomme Q
    6. Landsness EC
    7. Brichant JF
    8. Phillips C
    9. Massimini M
    10. Laureys S
    11. Tononi G
    12. Boly M

    (2011) Propofol anesthesia and sleep: a high-density EEG study

    Sleep 34:283–291.

    https://doi.org/10.3410/f.11117956.12356054

    • PubMed
    • Google Scholar
55. 1. Mölle M
    2. Marshall L
    3. Gais S
    4. Born J

    (2002)

    Grouping of spindle activity during slow oscillations in human non-rapid eye movement sleep

    Journal of Neuroscience 22:10941–10947.

    • PubMed
    • Google Scholar
56. 1. Müller B
    2. Gäbelein WD
    3. Schulz H

    (2006)

    A taxonomic analysis of sleep stages

    Sleep 29:967–974.

    • PubMed
    • Google Scholar
57. 1. Nelson LE
    2. Guo TZ
    3. Lu J
    4. Saper CB
    5. Franks NP
    6. Maze M

    (2002) The sedative component of anesthesia is mediated by GABA(A) receptors in an endogenous sleep pathway

    Nature Neuroscience 5:979–984.

    https://doi.org/10.1038/nn913

    • PubMed
    • Google Scholar
58. 1. Niedermeyer E
    2. da Silva FL

    (2005)

    Electroencephalography: Basic Principles, Clinical Applications, and Related Fields

    Electroencephalography: Basic Principles, Clinical Applications, and Related Fields, Lippincott Williams & Wilkins.

    • Google Scholar
59. 1. Niknazar M
    2. Krishnan GP
    3. Bazhenov M
    4. Mednick SC

    (2015) Coupling of thalamocortical sleep oscillations are important for memory consolidation in humans

    PLoS One 10:e0144720.

    https://doi.org/10.1371/journal.pone.0144720

    • PubMed
    • Google Scholar
60. 1. Nádasdy Z
    2. Hirase H
    3. Czurkó A
    4. Csicsvari J
    5. Buzsáki G

    (1999)

    Replay and time compression of recurring spike sequences in the hippocampus

    Journal of Neuroscience 19:9497–9507.

    • PubMed
    • Google Scholar
61. 1. Parrino L
    2. Terzano MG

    (1996) Polysomnographic effects of hypnotic drugs. A review

    Psychopharmacology 126:1–16.

    https://doi.org/10.1007/BF02246405

    • PubMed
    • Google Scholar
62. 1. Pennartz CM
    2. Lee E
    3. Verheul J
    4. Lipa P
    5. Barnes CA
    6. McNaughton BL

    (2004) The ventral striatum in off-line processing: ensemble reactivation during sleep and modulation by hippocampal ripples

    Journal of Neuroscience 24:6446–6456.

    https://doi.org/10.1523/JNEUROSCI.0575-04.2004

    • PubMed
    • Google Scholar
63. 1. Popa D
    2. Duvarci S
    3. Popescu AT
    4. Lena C
    5. Pare D

    (2010) Coherent amygdalocortical theta promotes fear memory consolidation during paradoxical sleep

    PNAS 107:6516–6519.

    https://doi.org/10.1073/pnas.0913016107

    • Google Scholar
64. 1. Prell GD
    2. Khandelwal JK
    3. Burns RS
    4. LeWitt PA
    5. Green JP

    (1988)

    Elevated levels of histamine metabolites in cerebrospinal fluid of aging, healthy humans

    Comprehensive Gerontology. Section A, Clinical and Laboratory Sciences 2:114–119.

    • PubMed
    • Google Scholar
65. 1. Purdon PL
    2. Pierce ET
    3. Mukamel EA
    4. Prerau MJ
    5. Walsh JL
    6. Wong KFK
    7. Salazar-Gomez AF
    8. Harrell PG
    9. Sampson AL
    10. Cimenser A
    11. Ching S
    12. Kopell NJ
    13. Tavares-Stoeckel C
    14. Habeeb K
    15. Merhar R
    16. Brown EN

    (2013) Electroencephalogram signatures of loss and recovery of consciousness from propofol

    PNAS 110:E1142–E1151.

    https://doi.org/10.1073/pnas.1221180110

    • Google Scholar
66. Book

    1. Rechtschaffen A
    2. Kales A

    (1968)

    A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects

    Bethesda, Md: U. S. National Institute of Neurological Diseases and Blindness, Neurological Information Network.

    • Google Scholar
67. 1. Robinson PA
    2. Phillips AJK
    3. Fulcher BD
    4. Puckeridge M
    5. Roberts JA

    (2011) Quantitative modelling of sleep dynamics

    Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 369:3840–3854.

    https://doi.org/10.1098/rsta.2011.0120

    • Google Scholar
68. 1. Rudolph M
    2. Pospischil M
    3. Timofeev I
    4. Destexhe A

    (2007) Inhibition determines membrane potential dynamics and controls action potential generation in awake and sleeping cat cortex

    Journal of Neuroscience 27:5280–5370.

    https://doi.org/10.1523/JNEUROSCI.4652-06.2007

    • PubMed
    • Google Scholar
69. 1. Sanchez-Vives MV
    2. McCormick DA

    (2000) Cellular and network mechanisms of rhythmic recurrent activity in neocortex

    Nature Neuroscience 3:1027–1034.

    https://doi.org/10.1038/79848

    • PubMed
    • Google Scholar
70. 1. Saunders A
    2. Granger AJ
    3. Sabatini BL

    (2015) Corelease of acetylcholine and GABA from cholinergic forebrain neurons

    eLife 4:e06412.

    https://doi.org/10.7554/eLife.06412

    • PubMed
    • Google Scholar
71. 1. Schellenberger Costa M
    2. Weigenand A
    3. Ngo HV
    4. Marshall L
    5. Born J
    6. Martinetz T
    7. Claussen JC

    (2016) A thalamocortical neural mass model of the EEG during NREM sleep and Its response to auditory stimulation

    PLOS Computational Biology 12:e1005022.

    https://doi.org/10.1371/journal.pcbi.1005022

    • PubMed
    • Google Scholar
72. 1. Sheroziya M
    2. Timofeev I

    (2014) Global intracellular slow-wave dynamics of the thalamocortical system

    Journal of Neuroscience 34:8875–8893.

    https://doi.org/10.1523/JNEUROSCI.4460-13.2014

    • PubMed
    • Google Scholar
73. 1. Silber MH
    2. Ancoli-Israel S
    3. Bonnet MH
    4. Chokroverty S
    5. Grigg-Damberger MM
    6. Hirshkowitz M
    7. Kapen S
    8. Keenan SA
    9. Kryger MH
    10. Penzel T
    11. Pressman MR
    12. Iber C

    (2007)

    The visual scoring of sleep in adults

    Journal of Clinical Sleep Medicine 3:121–131.

    • PubMed
    • Google Scholar
74. Book

    1. Sivagnanam S
    2. Majumdar A
    3. Yoshimoto K
    4. Astakhov V
    5. Bandrowski A
    6. Martone ME
    7. Carnevale NT

    (2013)

    Introducing the Neuroscience Gateway

    IWSG.

    • Google Scholar
75. 1. Steriade M
    2. Deschénes M

    (1984) The thalamus as a neuronal oscillator

    Brain Research Reviews 8:1–63.

    https://doi.org/10.1016/0165-0173(84)90017-1

    • Google Scholar
76. 1. Steriade M
    2. Deschênes M
    3. Domich L
    4. Mulle C

    (1985)

    Abolition of spindle oscillations in thalamic neurons disconnected from nucleus reticularis thalami

    Journal of Neurophysiology 54:1473–1497.

    • PubMed
    • Google Scholar
77. 1. Steriade M
    2. Dossi RC
    3. Nuñez A

    (1991)

    Network modulation of a slow intrinsic oscillation of cat thalamocortical neurons implicated in sleep delta waves: cortically induced synchronization and brainstem cholinergic suppression

    Journal of Neuroscience 11:3200–3217.

    • PubMed
    • Google Scholar
78. Book

    1. Steriade M
    2. Jones EG
    3. Llinas R

    (1990)

    Thalmic Oscillations and Signaling

    New York: John Wiley & Sons.

    • Google Scholar
79. 1. Steriade M
    2. Llinás RR

    (1988)

    The functional states of the thalamus and the associated neuronal interplay

    Physiological Reviews 68:649–742.

    • PubMed
    • Google Scholar
80. 1. Steriade M
    2. McCormick D
    3. Sejnowski T

    (1993a) Thalamocortical oscillations in the sleeping and aroused brain

    Science 262:679–685.

    https://doi.org/10.1126/science.8235588

    • Google Scholar
81. 1. Steriade M
    2. Nuñez A
    3. Amzica F

    (1993b)

    Intracellular analysis of relations between the slow (< 1 Hz) neocortical oscillation and other sleep rhythms of the electroencephalogram

    Journal of Neuroscience  13:3266–3283.

    • Google Scholar
82. 1. Steriade M
    2. Timofeev I
    3. Grenier F

    (2001)

    Natural waking and sleep states: a view from inside neocortical neurons

    Journal of Neurophysiology 85:1969–1985.

    • PubMed
    • Google Scholar
83. 1. Szymusiak R
    2. Alam N
    3. McGinty D

    (2000) Discharge patterns of neurons in cholinergic regions of the basal forebrain during waking and sleep

    Behavioural Brain Research 115:171–253.

    https://doi.org/10.1016/S0166-4328(00)00257-6

    • PubMed
    • Google Scholar
84. 1. Timofeev I
    2. Grenier F
    3. Bazhenov M
    4. Sejnowski TJ
    5. Steriade M

    (2000) Origin of slow cortical oscillations in deafferented cortical slabs

    Cerebral Cortex 10:1185–1199.

    https://doi.org/10.1093/cercor/10.12.1185

    • PubMed
    • Google Scholar
85. 1. Timofeev I
    2. Grenier F
    3. Steriade M

    (2001) Disfacilitation and active inhibition in the neocortex during the natural sleep-wake cycle: an intracellular study

    PNAS 98:1924–1929.

    https://doi.org/10.1073/pnas.041430398

    • PubMed
    • Google Scholar
86. 1. Timofeev I
    2. Steriade M

    (1996)

    Low-frequency rhythms in the thalamus of intact-cortex and decorticated cats

    Journal of Neurophysiology 76:4152–4168.

    • PubMed
    • Google Scholar
87. 1. Vanini G
    2. Lydic R
    3. Baghdoyan HA

    (2012) GABA-to-ACh ratio in basal forebrain and cerebral cortex varies significantly during sleep

    Sleep 35:1325–1334.

    https://doi.org/10.5665/sleep.2106

    • PubMed
    • Google Scholar
88. 1. Vanini G
    2. Wathen BL
    3. Lydic R
    4. Baghdoyan HA

    (2011) Endogenous GABA levels in the pontine reticular formation are greater during wakefulness than during rapid eye movement sleep

    Journal of Neuroscience 31:2649–2656.

    https://doi.org/10.1523/JNEUROSCI.5674-10.2011

    • PubMed
    • Google Scholar
89. 1. Vazquez J
    2. Baghdoyan HA

    (2001)

    Basal forebrain acetylcholine release during REM sleep is significantly greater than during waking

    American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 280:R598–601.

    • PubMed
    • Google Scholar
90. 1. von Krosigk M
    2. Bal T
    3. McCormick DA

    (1993) Cellular mechanisms of a synchronized oscillation in the thalamus

    Science 261:361–364.

    https://doi.org/10.1126/science.8392750

    • PubMed
    • Google Scholar
91. 1. Wei Y
    2. Krishnan GP
    3. Bazhenov M

    (2016) Synaptic mechanisms of memory consolidation during sleep slow oscillations

    Journal of Neuroscience 36:4231–4247.

    https://doi.org/10.1523/JNEUROSCI.3648-15.2016

    • PubMed
    • Google Scholar
92. 1. Werth E
    2. Achermann P
    3. Borbély AA

    (1996) Brain topography of the human sleep EEG: antero-posterior shifts of spectral power

    Neuroreport 8:123–127.

    https://doi.org/10.1097/00001756-199612200-00025

    • PubMed
    • Google Scholar
93. 1. Wulff K
    2. Gatti S
    3. Wettstein JG
    4. Foster RG

    (2010) Sleep and circadian rhythm disruption in psychiatric and neurodegenerative disease

    Nature Reviews Neuroscience 11:589–599.

    https://doi.org/10.1038/nrn2868

    • PubMed
    • Google Scholar
94. 1. Yu X
    2. Ye Z
    3. Houston CM
    4. Zecharia AY
    5. Ma Y
    6. Zhang Z
    7. Uygun DS
    8. Parker S
    9. Vyssotski AL
    10. Yustos R
    11. Franks NP
    12. Brickley SG
    13. Wisden W
    14. Ma Y ZZ

    (2015) Wakefulness Is governed by GABA and histamine cotransmission

    Neuron 87:164–178.

    https://doi.org/10.1016/j.neuron.2015.06.003

    • PubMed
    • Google Scholar

Author details

1. Giri P Krishnan

   Department of Medicine, University of California, San Diego, La Jolla, CA, United States

   Contribution

   GPK, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   For correspondence

   gkrishnan@ucsd.edu

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0002-3931-7633

2. Sylvain Chauvette

   1. Department of Psychiatry and Neuroscience, Université Laval, Québec, Canada
   2. Centre de Recherche de l’Institut Universitaire en Santé Mentale de Québec, Université Laval, Québec, Canada

   Contribution

   SC, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

3. Isaac Shamie

   Departments of Radiology and Neurosciences, University of California, San Diego, La Jolla, CA, United States

   Contribution

   IS, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

4. Sara Soltani

   1. Department of Psychiatry and Neuroscience, Université Laval, Québec, Canada
   2. Centre de Recherche de l’Institut Universitaire en Santé Mentale de Québec, Université Laval, Québec, Canada

   Contribution

   SS, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

5. Igor Timofeev

   1. Department of Psychiatry and Neuroscience, Université Laval, Québec, Canada
   2. Centre de Recherche de l’Institut Universitaire en Santé Mentale de Québec, Université Laval, Québec, Canada

   Contribution

   IT, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

6. Sydney S Cash

   Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, United States

   Contribution

   SSC, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

7. Eric Halgren

   Departments of Radiology and Neurosciences, University of California, San Diego, La Jolla, CA, United States

   Contribution

   EH, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

8. Maxim Bazhenov

   Department of Medicine, University of California, San Diego, La Jolla, CA, United States

   Contribution

   MB, Conception and design, Analysis and interpretation of data, Drafting or revising the article

   Competing interests

   The authors declare that no competing interests exist.

• 3,630

  views

• 688

  downloads

• 69

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.