Non-REM (Thalamocortical) Sleep DE Moul, Cleveland Clinic, Cleveland, OH, USA r 2014 Elsevier Inc. All rights reserved. This article is a revision of the previous edition article by M Steriade, volume 3, pp 640–644, r 2003, Elsevier Inc.
Introduction Nonrapid eye movement (non-REM) sleep is one of the general modes (i.e., wake, REM sleep, TCS) of consciousness. A better name is ‘thalamocortical sleep’ (TCS).
Clinical Scoring TCS is now scored by 30-s epochs into the ordinal stages of N1, N2, and N3 on the polysomnogram (PSG). During full wakefulness and rapid eye movement (REM) sleep, the electroencephalogram (EEG) channels of the PSG exhibit fast, mixed-frequency waveforms. By comparison, the obvious waveforms of TCS are slower and of higher voltage. In relaxed wakefulness, a (8–12 Hz) waveforms occur, and, as they disappear, their absence and y (4–7 Hz) waveforms mark the beginning of Stage N1. K complexes (KCs) and sleep spindles are specific transient waveforms distinctive to Stages N2 and N3. KCs appear as a brief surface-negative phase followed by a slow (o0.5 Hz) positive phase. A sleep spindle is a 12–14 Hz waveform lasting 0.5 s or longer, which has a crescendo/ decrescendo resembling a spindle. d Waves (0.5–4 Hz) later come to be of higher voltage, and when they constitute more than 20% of an epoch, Stage N3 is scored.
Physiological Regulation ‘Thalamocortical sleep’ should not be taken to mean that only TCS-specific processes govern TCS occurrence. Other wholebrain processes regulate the modes of wakefulness, REM sleep, and TCS. The sleep cycle process alternates TCS and REM sleep during nighttime sleep. The timing and intensity of wake, REM sleep, and TCS is governed by TCS-homeostatic, REMhomeostatic, circadian, infraradian, hormonal, and seasonal processes. Other background influences also govern the kind of TCS that occurs, including a person’s age and gender. Respiratory and other functions are contingent on the specific form of TCS (e.g., Stage N1 vs. Stage N3), and they in turn have feedback effects on TCS itself (e.g., TCS characteristics affect risks of sleep apnea events, but the apneic events then in turn affect TCS).
Distinctive Pattern of Synchronous Discharge Neural networks during TCS have patterns of masssynchronized discharging. Neurons synchronously hyperpolarize and then depolarize in burst-mode action potentials across the neural network. A visual analogy is that of ‘waves’ of standing and sitting that crowds engage in at sporting events: individuals acting synchronously in massed unison
Encyclopedia of the Neurological Sciences, Volume 3
give rise to macroscopic waves. During TCS, the dynamic electrical behavior of the thalamocortical network produces the large-scale potentials that can be detected with EEG scalp electrodes. These macroscopic waves are synchronous but not seizures. Seizures are also synchronous but have distinctive waveforms. Unlike during seizures, during TCS the sleeper remains partially conscious of the environment. Nonetheless, TCS synchrony can foster seizures. For this reason, intentional sleep deprivation is used as a facilitator to uncover seizures for their diagnosis, whereas causes of sleep deprivation are treated in patients with epilepsy to avoid seizures. During TCS, neuronal membranes are generally more hyperpolarized than during wakefulness or REM sleep. This average hyperpolarization is made possible when wakeassociated depolarizing signals from the brainstem via the ascending reticular activating system (ARAS) are muffled at TCS onset. When this hyperpolarization regimen occurs, neurons can take on a pattern of rhythmic burst firing that is triggered by hyperpolarization-triggered slow-depolarizing calcium currents. ‘Hyperpolarization-triggered depolarization’ may sound paradoxical, but it is like the tick-tock gating of pendulum movements in mechanical clocks. While on time average, the neurons are more hyperpolarized, rhythmic, burst-mode depolarizations of short duration occur during TCS. This burst-mode discharging is ultimately the basis for observed EEG waveforms. With complex feedforward and feedback signaling during TCS along complex wiring pathways (Figure 1) between the thalamus and neocortex, synchronized macroscopic EEG waveforms appear that are slower than during wakefulness or REM sleep but occur in chaotic sequences. During TCS, the EEG waves are synchronous both across neocortical hemispheres and anterioposteriorily. This synchrony is made possible by the circuitry between the neurons of the reticular (RE) thalamic nucleus; the thalamocortical glomerulus composed of the thalamocortical neuron (ThCx) and its supporting inhibitory interneurons utilizing gaminobutyric acid (GABA) as neurotransmitter; the organizational configuration of the somatosensory neocortex (Cx); and their feedback connections, as cartooned in Figure 1. Drugs facilitating chloride currents through GABAergic channels (e.g., ethanol and benzodiazepine receptor agonists) generally bias the network toward average hyperpolarization, making burst-mode discharging more likely. During wakefulness, many such drugs are anxiolytics; during TCS they also change the EEG and foster TCS. Although drugs can electrically bias neuronal membranes, the circuit as a whole does not rigidly predetermine specific wave appearances over long time periods because normally the global circuit behaves with too much complexity to allow rigid, linear determinism. Such determinism would preclude responsiveness to the environment. Instead, PSG waveforms behave like weather storms. One can predict general storm
Non-REM (Thalamocortical) Sleep
KC + spindle Cx Cx b−c c
KC + -ThCx Cx c
KC + - Cx Cx d′
Figure 1 Depiction of postulated EEG waveform trains observed under differing thalamocortical circuit connectivities. (A) Simple circular dependency. KC arising at the cortex generates timing effect (a) from the Cx to RE that time-synchronizes sleep spindle RE burst-mode firing (left lower tracing, seen arising from slower depolarizing potentials). In turn, RE hyperpolarizes (b) ThCx to push it into burst-mode firing as well. Completing the circle, The ThCx burst-mode firing stimulates Cx (c) to have repetitive mass discharges seen on the surface EEG channel as a spindle following immediately after the KC. (B) With similar cell actions to (A), but where Cx has more or less simultaneous effect on RE and ThCx, but with EEG showing KC followed by delta-frequency waves. (C) Here there are Cx–Cx connectivities that tend to produce a similar-appearing waveform to (B), but now through another separate set of circuit connectivities. Adapted from Amzica F and Steriade M (2002) The functional significance of K-complexes. Sleep Medicine Reviews 6: 139–149.
patterns but not the location of particular lightning strikes. Likewise, one can predict general TCS waveform rates but not the timing of individual waveforms. In Figure 1, Amzica and Steriade portrayed three scenarios of interactions between the ThCx, Cx, and RE cells.
Two Key Forces Driving TCS Although chaotic in sequence, the waveforms have two main driving forces during TCS. The first force arises from the RE, whose neurons have three salient characteristics. First, they
Non-REM (Thalamocortical) Sleep
are GABAergic, so that their synaptic transmission drives hyperpolarization (GABA: g-aminobutyric acid). Second, they have dendrodentric inhibitory synapses between them, so that when some RE cells become hyperpolarized and begin burst discharging, the immediacy of communication between them synchronizes all their burst-mode discharges. Because their efferents globally innervate the somatosensory thalamus bilaterally, they are configured to drive globally synchronized ThCx burst-mode discharging in turn. Third, they hyperpolarize more easily than ThCxs. With sleep onset, RE neurons are the first to start burst-mode discharging, so they are a driving force for TCS synchronization. The sleep spindle has been demonstrated to derive directly from this mechanism. Unlike RE neurons, ThCxs are glutaminergic and hence are generally depolarizing in their direct postsynaptic effects in the neocortex. However, on the hyperpolarized regimen of the neocortex during stable TCS, such ThCx discharges probably have more timing-of-discharge effects than they do an overall depolarizing effect. At times this balance may sway more in the direction of depolarization and induce a stochastically normal ‘microarousal’ during TCS (observed as a short period of higher EEG frequencies). An additional complexity is that glutaminergic postsynaptic receptors can be ionotropic or metabotropic. The former leads to more prompt signaling effects, whereas the latter may take minutes before the second-messenger cascade in the postsynaptic neuron results in change in discharge characteristics. Because the postsynaptic neocortical field of the ThCx includes neurons of many types (some excitatory and some inhibitory), it is easy to appreciate why TCS waveforms have a chaotic sequence. A second driving force conditioning TCS waveforms originates from the neocortex itself. In vitro, neocortical slices produce a slow synchronous (o0.5 Hz) rhythm. Its cyclic phases are termed ‘UP’ and ‘DOWN.’ The UP phase is a time period when neocortical neurons produce action potentials whereas the DOWN phase is a period of relative electrical silence. This slow rhythm has feedforward effects on thalamic discharges. For most of the TCS record, this slow rhythm is not visible. However, at times these slow waves may be amplified by instantaneous network events into EEG visibility. The slow rhythm is thought to be integral to the formation of KCs. It has long been appreciated that KCs during Stage N2 sleep can occur in response to an environmental stimulus. Legend has it that the ‘K’ stands for ‘Knock,’ from the longknown observation that some KCs can be induced by knocking sounds in the sleeper’s environment. For decades the puzzle was why a KC, which is a stereotypical marker for Stage N2 sleep, and reflective of deeper and more stable sleep than Stage N1 sleep, should nonetheless be inducible by an environmental stimulus, and thereby resemble a microarousal. If a ‘microarousal’ is part of stable sleep, then why would it be part of ‘nonconscious’ sleep?
Rethinking TCS for What It Is, Not for What It Is Not The traditional concept of sleep overemphasized the contrast between sleep and full wakefulness. Wakefulness has long been considered by opinion leaders as the only relevant domain of
life. For example, Jean-Paul Sartre wrote ‘‘Consciousness is only consciousness of something,’’ as if only a condition of highly focused thought could count as consciousness. For many centuries sleep has been thought of as void of sense and worth. Dreaming and REM sleep have enjoyed a fairer treatment since Freud promoted dream life as a focus of inquiry, but TCS, as the ‘Non-REM sleep’ name suggests, continues to be stigmatized. This bias gets in the way of understanding TCS for what it is, and what it is not. If one accepts the dogma that, because fast EEG activity is typical of wakefulness, then healthy TCS would only be ‘healthy’ when all traces of EEG fast activity are absent. TCS would then be thought of dogmatically as akin to coma, as merely a kind of ‘deadness to the world.’ But KCs do not conform to this dogma, as KCs are paradoxically ‘partial arousals’ that also are ‘markers of sleep.’ The traditional dogma is not helpful. If one instead considers TCS as a mode of consciousness, then EEG findings like the KCs become easier to comprehend: They are features of a normal mode of consciousness, as a composite of neural network events sketched in Figure 1 that we do not yet fully understand. The ‘nonconsciousness’ of TCS is partially true, in the limited sense that with the onset of TCS there develops a muffling of sensory input. When called a ‘stimulus barrier’ it is misnamed insofar as it suggests an impermeable wall against all environmental stimuli. The sensory muffling that occurs is a disfacilitation of the ARAS, but not an exhaustive expungement, of all sensory signaling. Even though muffled, some sensory input can move from peripheral sensory organs through the brainstem to the thalamus to neocortical endpoints. This transmission can be observed in sensory evoked potentials. For a KC to be responsive to environmental stimuli, the ‘stimulus barrier’ cannot be absolute. Some sensory signal strength from peripheral inputs does arrive at neocortical sites during TCS. Sensory signals from such stimuli as a knocking sound, arriving at the neocortex at flashpoints when UP slow waves are prepotent, would lead to amplification of slow wave mass discharges throughout the neocortex, and then to the thalamus in a feedforward fashion. Such a discharge would be observable on the EEG as a KC across wide regions of the somatosensory neocortex. Not all knocks produce KCs, as not all sensory stimuli may happen to arrive at the neocortex at opportune flashpoints in the neocortical slow rhythm. Yet curiously, some KCs yield more prolonged microarousals, whereas others yield none at all, or may be associated with entrained sleep spindles. For KCs that yield no ‘microarousals,’ the network-wide effect of a mass discharge from glutaminergic ThCxs may only relate to a timing effect, rather than an additional general depolarization effect. In this case, KC’s effect is merely to reverberate burst discharging within an overall hyperpolarization regimen in the neocortex. For other KCs, the membrane potentials in the ThCx with a KC discharge may be optimal, by happenstance, to foster a sleep spindle driven by the RE immediately thereafter. Several d-frequency waves may also appear, and are ambiguous as to whether they represent more or less resemblance to full wakefulness, or deeper sleep, in their overall physiological effects. For yet other KCs, the discharge may occur under thalamocortical network conditions, wherein the ThCx discharges actually move the network briefly into a depolarized regimen, leading to trailing fast EEG waveforms typical of microarousals. Sometimes KCs might actually awaken the sleeper. Furthermore, some KCs
Non-REM (Thalamocortical) Sleep
may not actually have any sensory triggers but arise as ‘rogue waves’ within TCS. Prompted and unprompted awakenings are normal in a bout of TCS.
The Broader Ontology of TCS Controversies about TCS have included findings about observed wave packets associated with daytime symptoms of poor sleeping. The phenomena of cyclic alternating pattern (CAP) and its subtypes, as associated with cases of insomnia, bruxism, and adverse sleeping more generally, may represent modes of TCS where the relative dynamic temporal event trains of hyperpolarization and depolarization lead to self-reported adverse consequences. Likewise, the alpha–delta pattern observed during Stage N3 may represent miscoordinated TCS, and give rise to daytime symptoms. The phenomena of CAP, alpha–delta pattern, and KCs all appear to be part of a broader, heterogeneous ontology of TCS than conventional scoring rules would suggest. Two mysteries about TCS endure within sleep medicine. One concerns why some individuals with severe sleep apnea, in whom TCS is relentlessly fragmented, do not appear to have any daytime neurocognitive sequelae, whereas others with virtually the same PSG findings have terrible sleepiness and mental dysfunction. The other mystery is why some individuals with paradoxical insomnia have apparently normal TCS on the PSG but nonetheless complain bitterly that they did not sleep at all, or have had very poor sleep. In pondering these mysteries, it is important to remember that TCS observed on the PSG is only a surface representation of brain activity. The electrical potentials seen with EEG leads reflect only the electrical potentials as recorded from the thin depth of neocortex nearest to the EEG scalp electrodes and do not necessarily report about events in deep brain structures. Deeper structures are where key regulatory nuclei for neurocognitive functioning and self-perception may reside. An illustrative fact is that limbic cortex, which is not available for recording in the EEG, does not have the same connections to thalamic nuclei that the general somatosensory cortex does. If limbic or other deep structures are involved with one’s sense of sleep, then conclusions drawn from the PSG concerning the adequacy or inadequacy of TCS may be at least partly mistaken. A person’s sense of alertness or restoration derived from TCS may not correlate well with PSG findings. In this respect, the phenomenological ontology of TCS appears to include kinds of TCS that current neuroscience has yet to investigate. Some normally functioning persons require only short (B4 h) total daily sleep, whereas others require uncommonly long (B10–11 h) total daily sleep. Human TCS normally occupies approximately 80% of total sleep, so for short and long sleepers, the kinds of TCS must differ. Proposed functions for TCS have included support of general anabolic cellular processes, synaptic pruning, mnemonic stabilizing of daytime-acquired information, bodily repair, and growth. The existence of short and long sleepers suggests that the ontology of TCS needs to be understood broadly, to account for normal interindividual variability in nondiseased TCS. Such an ontology of normal individual differences in TCS pose ethical and public health policy implications about
an individual’s particular work duties, should the work requirements obstruct the specific TCS that normal individuals may variably require. Additional findings suggest that TCS is a mode of consciousness. Steriade and colleagues have clearly shown that there are fast and coherent g (B40 Hz) waveforms that occur during TCS. This finding belies the notion that TCS waveforms should always be slow, and suggests that something like thinking is occurring during TCS. More recent literature points to the separable memory enhancements from TCS and REM sleep. Furthermore, experimental work in animals and humans has documented that regions of neocortex that were more highly utilized during wakefulness show persistent EEG changes, and metabolic effects, during TCS on the night after the test conditions. These several phenomena have led to a growing consensus that there are region-specific events occurring during TCS. The clinical implications of local sleep effects are quite broad and raise the issue of the collection of TCS parasomnias, when the sleeper is ‘half awake.’ An example is the confusional arousal in which a person has a partial, confused awakening. Furthermore, because one typically enters into sleep through TCS, there are also TCS-onset phenomena, such as the common sleep start. Both the parasomnias and sleep onset phenomena involve TCS.
Avoiding Mereological Fallacies When Thinking About TCS Established sleep staging criteria have been the terms in which sleep architecture has been understood for half a century. Now newer TCS neurobiology suggests that hidden mereological (part-to-whole) fallacies may have inhibited deeper understanding. The EEG phenomena observed during TCS is only part of the story about TCS considered as a whole. TCS can occur in some parts of the brain but not necessarily in every part of the brain simultaneously. Some kinds of TCS are not psychologically restorative, whereas other forms are. The existence of TCS variants encourages TCS to be accepted as a wide-ranging, ontologically complex mode of consciousness.
See also: Action Potential, Generation of. Aging and Sleep. Ascending Reticular Activating System (ARAS). Coma. Consciousness. Dendrites. Drowsiness. Electroencephalogram (EEG). Electroencephalographic Spikes and Sharp Waves. Epileptogenesis. Ethanol. Ethical Issues. Event-Related Potentials (ERPs). Excessive Daytime Sleepiness. GABAA Receptor Channels; Properties and Regulation. Hypnotics. Ion Channels, Overview. Membrane Potential. Memory, Overview. Parasomnias. Polysomnography, Clinical. Polysomnography; Technique. Sleep and Epilepsy. Sleep, Condition of. Sleep Disorders. Sleep; Overview. Sleep–Wake Cycle. Wakefulness
Further Reading American Academy of Sleep Medicine (2005) The International Classification of Sleep Disorders: Diagnostic and Coding Manual, 2nd edn. Westchester, IL: American Academy of Sleep Medicine.
Non-REM (Thalamocortical) Sleep
Amzica F and Steriade M (2002) The functional significance of K-complexes. Sleep Medicine Reviews 6: 139–149. Colrain IM (2005) The K-complex: A 7-decade history. Sleep 28: 255–273. Crunelli V and Hughes SW (2010) The slow (o1 Hz) rhythm of non-REM sleep: A dialogue between three cardinal oscillators. Nature Neuroscience 13: 9–17. Destexhe A, Hughes SW, Rudolph M, and Crunelli V (2007) Are corticothalamic ’up’ states fragments of wakefulness? Trends in Neuroscience 30: 334–342. Foldvary-Schaefer N, Krishna J, Budur K, and Cleveland Clinic Sleep Disorders Center (2010) A Case a Week: Sleep Disorders by the Cleveland Clinic. Oxford; New York: Oxford University Press. Iber C and American Academy of Sleep Medicine (2007) The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology, and Technical Specifications. Westchester, IL: American Academy of Sleep Medicine. Kryger MH, Roth T, and Dement WC (2011) Principles and Practice of Sleep Medicine, 5th edn. Philadelphia, PA: Saunders/Elsevier.
Le Van Quyen M and Bragin A (2007) Analysis of dynamic brain oscillations: methodological advances. Trends in Neuroscience 30: 365–373. Ogilvie RD and Harsh JR (1994) Sleep Onset: Normal and Abnormal Processes. Washington, DC: American Psychological Association. Steriade M and McCarley R (2005) Brainstem Control of Wakefulness and Sleep, 2nd edn. New York: Plenum Press. Terzano MG, Parrino L, Sherieri A, et al. (2001) Atlas, rules, and recording techniques for the scoring of cyclic alternating pattern (CAP) in human sleep. Sleep Medicine 2: 537–553. Terzano MG, Parrino L, Sherieri A, et al. (2002) Atlas, rules, and recording techniques for the scoring of cyclic alternating pattern (CAP) in human sleep. Sleep Medicine 3: 187–199. Tononi G (2009) Slow wave homeostasis and synaptic plasticity. Journal of Clinical Sleep Medicine 5(supplement 2): S16–S19.