Sleep Disorders

Disorders are dysregulations of the architecture

Module 19 established that normal sleep is the orderly cycling of distinct neuromodulatory states, each with characteristic graphoelements and each held within its proper place in the night by active switching circuits. The major sleep disorders are best understood not as random aberrations but as failures of state-boundary control: a state intruding where it does not belong, a state fragmenting before it can consolidate, or a single defining feature of a state - most strikingly REM atonia - decoupling from the rest of that state and appearing or disappearing on its own. This framing collapses a long differential diagnosis into a small set of mechanistic questions. For any nocturnal phenomenon, ask: which state is involved, what boundary has failed, and which polysomnographic channel reveals the failure? Answer those three and the diagnosis usually follows.

The diagnostic toolkit accordingly extends well beyond the routine EEG. Polysomnography (PSG) adds respiratory effort belts (thoracic and abdominal), nasal pressure and thermal airflow, pulse oximetry, electrocardiography, body-position sensing, and anterior tibialis (limb) EMG to the EEG-EOG-chin-EMG montage of Module 19, so that the breath, the heart, the limbs, and the brain are recorded in register. The Multiple Sleep Latency Test (MSLT) quantifies daytime sleep propensity and the speed of REM onset across a structured series of daytime naps. Actigraphy estimates the rest-activity pattern over weeks, and the clinical history - witnessed events, timing within the night, dream recall, family history - frames every instrumented finding. Throughout, the EEG remains the substrate that defines sleep itself, so the staging fluency built in Module 19 is precisely what makes these elaborate studies interpretable. A respiratory event matters because of the arousal it produces, and an arousal is an EEG judgment.

Narcolepsy and the sleep-onset REM period

Narcolepsy is the prototypical disorder of REM state instability and of the wake-promoting system that normally stabilizes all state boundaries. In narcolepsy type 1, the underlying lesion is a profound, selective loss of the roughly tens of thousands of hypothalamic orexin (hypocretin) neurons of the lateral hypothalamus, widely thought to be autoimmune in origin and strongly associated with the HLA-DQB106:02 allele. Orexin is the principal stabilizer of wakefulness: it provides excitatory tone to the aminergic and cholinergic arousal systems and, crucially, to the REM-off side of the brainstem flip-flop described in Module 19. Remove it and the switches that hold wake and REM apart become unstable, so the boundaries between wake, NREM, and REM grow porous and REM phenomena erupt at inappropriate times. This orexin deficiency is directly measurable: under current ICSD-3 criteria, narcolepsy type 1 can be established by typical cataplexy together with either the polysomnographic-MSLT findings below or a low cerebrospinal-fluid hypocretin-1 (orexin-A) concentration - at or below 110 pg/mL, or less than one third of mean normal values - which is the single most specific diagnostic test and stands in for the MSLT when it is available. The clinical tetrad follows directly: excessive daytime sleepiness (the wake state cannot be held), cataplexy (REM atonia intruding abruptly into wakefulness, classically triggered by strong emotion such as laughter), sleep paralysis (REM atonia persisting into the wake transition at sleep onset or offset), and hypnagogic and hypnopompic hallucinations (REM dream imagery intruding into the wake-sleep transition). Narcolepsy type 2* shares the sleepiness and the supporting MSLT electrophysiology but lacks cataplexy and shows CSF hypocretin-1 that is normal (above 110 pg/mL) or not measured, and it is the less stable and less well-defined of the two diagnoses, with a meaningful rate of reclassification over time.

The electrophysiological signature is the sleep-onset REM period (SOREMP): the appearance of REM sleep within 15 minutes of sleep onset, which is abnormal because healthy adults interpose roughly 90 minutes of NREM before their first REM period (recall the normal REM latency of Module 19). The MSLT operationalizes the search for SOREMPs across five scheduled nap opportunities at two-hour intervals during the day, performed only after an adequate, PSG-documented night of sleep the night before. The supportive criteria for narcolepsy are a mean sleep latency of 8 minutes or less together with two or more SOREMPs across the test; by current AASM rules a SOREMP occurring within 15 minutes of sleep onset on the preceding night's PSG may substitute for one of the two required nap SOREMPs. Because the test is exquisitely sensitive to context, it is invalidated by insufficient prior sleep, recent shift work or jet lag, untreated obstructive sleep apnea, and REM-suppressing or REM-rebound-producing drugs. Antidepressants in particular - which suppress REM and, on withdrawal, produce REM rebound - must usually be tapered off well in advance, on the order of two weeks (longer for long-half-life agents such as fluoxetine), or the test cannot be trusted.

NREM parasomnias - the disorders of arousal

The NREM parasomnias - confusional arousals, sleepwalking (somnambulism), and sleep terrors - are the disorders of arousal, in which a partial awakening from slow-wave sleep leaves motor and emotional systems active while the systems that support consciousness, memory encoding, and executive judgment remain offline. They therefore arise predominantly from N3 in the first third of the night, when slow-wave pressure peaks, and are typified by amnesia for the event, mental confusion and slowed responsiveness if the person is awoken during it, automatic and often clumsy behavior, and a strong familial and childhood predisposition that usually wanes with age. The unifying mechanism is incomplete state dissociation: functional imaging and intracranial recordings of such events show that motor, cingulate, and limbic networks transition toward a waking pattern while frontoparietal association cortex and thalamocortical arousal systems remain in the local slow oscillations of deep NREM. The person is, quite literally, awake in some brain regions and asleep in others.

On the PSG the cardinal finding is an abrupt arousal out of N3, frequently preceded by a burst of high-amplitude rhythmic or hypersynchronous delta, with autonomic activation (tachycardia, increased respiration) in sleep terrors, and - critically - no evolving epileptiform discharge before or during the event. The behaviors emerge directly from deep sleep without an intervening period of full wakefulness, which is exactly what distinguishes them from the dream-enactment of REM behavior disorder, where the person rises out of REM. This absence of an ictal pattern is also the crux of the single most important and most frequently mishandled differential in all of nocturnal-event medicine.

Distinguishing parasomnias from nocturnal seizures

Sleep-related hypermotor epilepsy (SHE), formerly called nocturnal frontal lobe epilepsy, is the great mimic of NREM parasomnia, and getting the distinction right changes management completely - antiseizure medication and an epilepsy work-up on one side, reassurance and safety counseling on the other. The discriminators are largely electroclinical rather than purely electrographic. Seizures are typically brief, lasting seconds to about two minutes; highly stereotyped from night to night, so that the patient repeats nearly the same sequence; frequent, sometimes many times per night, and often clustered; able to occur at any point in the night, including light NREM and even relaxed wakefulness, rather than being locked to early-night N3; and characterized by dystonic or hyperkinetic posturing, asymmetric tonic postures, or vigorous repetitive movements. NREM parasomnias, by contrast, tend to be longer, far more variable in their semiology from event to event, infrequent, and tied to early-night slow-wave sleep. The EEG is decisive in principle - a seizure shows an ictal rhythm that evolves in frequency and spatial field over its course, as developed in the seizure pillar - but frontal and especially mesial or orbitofrontal seizures often have a scalp-negative ictal EEG or one buried under movement artifact, so a normal interictal study never excludes epilepsy. Capturing multiple events on video-EEG is sometimes the only reliable arbiter, and even then expert disagreement occurs.

Obstructive sleep apnea and the arousal as the EEG currency

In obstructive sleep apnea (OSA) the upper airway repeatedly narrows or collapses during sleep despite continued respiratory effort, and although the primary events are respiratory and mechanical, their decisive EEG consequence is the cortical arousal that terminates each event and restores airway tone. The AASM event definitions tie the respiratory and cortical worlds together. An apnea is a drop in airflow of at least 90 percent from baseline for at least 10 seconds; it is classified obstructive when respiratory effort persists throughout, central when effort is absent, and mixed when it begins central and becomes obstructive. A hypopnea is a lesser airflow reduction - by the recommended rule, at least a 30 percent drop for at least 10 seconds - accompanied by either a 3 percent or greater oxygen desaturation or an EEG arousal (an alternative, more conservative rule requires a 4 percent desaturation and does not count arousals, and which rule a laboratory uses materially changes the resulting index). The arousal itself is a purely EEG-defined event: an abrupt shift to faster EEG frequencies - alpha, theta, or frequencies above 16 Hz, but explicitly not spindles - lasting at least 3 seconds and preceded by at least 10 seconds of stable sleep; during REM, scoring an arousal additionally requires a concurrent increase in chin EMG amplitude lasting at least one second, because the REM EEG is already wake-like and cannot signal arousal on its own.

The summary metric, the apnea-hypopnea index (AHI) in events per hour of sleep, grades severity (commonly mild 5-15, moderate 15-30, severe above 30), but the EEG-defined arousal index is frequently what explains the clinical phenotype, and the two can diverge widely. The architectural cost of OSA is the heart of the disease. Each respiratory event terminates in a microarousal that fragments sleep, repeatedly demoting the patient back toward N1 and aborting descents into consolidated N3 and sustained REM, so that the patient cycles incessantly through light sleep without accumulating the deep and REM sleep that restore daytime function. The hypnogram of severe OSA is shredded: increased N1 and wake-after-sleep-onset, suppressed slow-wave and REM sleep, and an oximetry channel that saws up and down in lockstep with the respiratory events. This fragmentation, not the hypoxemia alone, is what drives much of the daytime sleepiness, and it is why a patient can spend eight hours in bed and remain profoundly unrefreshed. It is also exactly why untreated OSA must be excluded before a narcolepsy work-up is believed, since OSA-driven fragmentation and REM rebound can manufacture the short latencies and even the SOREMPs that the MSLT is designed to detect.

REM sleep behavior disorder - atonia decoupled from REM

REM sleep behavior disorder (RBD) is the cleanest example anywhere in sleep medicine of a single state feature failing in isolation. The REM-atonia circuit introduced in Module 19 - glutamatergic subcoeruleus neurons driving glycinergic and GABAergic premotor neurons in the ventromedial medulla, which hyperpolarize spinal and cranial motor neurons - fails, so the active muscle paralysis that normally accompanies REM is lost while the rest of REM, including vivid and often unpleasant or violent dream mentation, proceeds intact. The patient therefore enacts dreams: punching, kicking, gesturing, talking, shouting, and at times leaping from the bed, typically in the second half of the night when REM predominates (consistent with the back-loading of REM described in Module 19), usually with detailed dream recall on waking that is congruent with the observed behavior, and with a substantial risk of injury to the patient or the bed partner that frequently prompts the clinical visit.

The required diagnostic finding on PSG is REM sleep without atonia (RSWA, also abbreviated RWA) - the persistence of abnormally elevated tonic chin EMG tone and/or excessive phasic muscle twitching in the chin and limb leads during epochs that are otherwise unmistakably REM by their EEG and rapid eye movements. Quantitative criteria formalize the judgment: validated thresholds express, for example, the proportion of REM scored with sustained tonic chin activity and the density of short phasic bursts in combined chin and limb channels (the Montreal and SINBAR methods are the commonly cited quantitative approaches), and the diagnosis of clinical RBD requires both this electrophysiological RSWA and a history or video documentation of dream-enactment behavior. Synchronized video capturing the behavior strengthens the diagnosis and helps separate it from the periodic limb movements, apnea-related movements, and pseudo-RBD of severe untreated OSA, which can mimic enactment and must be excluded - sometimes by re-studying the patient on positive airway pressure. The interpreter's essential task is to recognize a dissociation that should never occur: the EEG and EOG declare REM while the EMG declares wakefulness, and that decoupling is itself the diagnosis.

A unifying view and the failure modes of interpretation

Step back and the four disorders array neatly along the boundary-failure axis. Narcolepsy is loss of the wake-stabilizing orexin signal, so REM elements intrude into wake and the wake-NREM-REM boundaries blur in both directions. NREM parasomnias are incomplete arousals from N3, in which motor and limbic systems wake while association cortex stays asleep. OSA is repetitive forced arousal that prevents any state from consolidating, fragmenting the architecture from the outside. RBD is the selective failure of one component - REM atonia - while the rest of REM proceeds. Each has a single channel that betrays it: the MSLT clock and SOREMPs for narcolepsy, the arousal-out-of-delta for NREM parasomnia, the arousal index and respiratory channels for OSA, and the chin and limb EMG for RBD. The discipline that ties the pillar together is that the EEG defines the state and therefore defines every one of these events - an arousal, a SOREMP, an awakening from N3, and the REM that RBD corrupts are all, at bottom, EEG and EMG judgments grounded in the staging of Module 19.

Several failure modes recur often enough to name explicitly. Trusting an MSLT without auditing the night and the medication list is the commonest narcolepsy error and manufactures false positives from sleep deprivation, apnea, and antidepressant withdrawal. Calling a stereotyped, brief, clustered nocturnal event a parasomnia because a single interictal EEG looked normal misses sleep-related hypermotor epilepsy, since frontal ictal patterns are often scalp-negative and demand event capture. Reading pseudo-RBD in severe OSA as true RBD attaches a grave prodromal-neurodegeneration label to what is really airway-driven movement, and the remedy is to treat the apnea and re-examine atonia. Letting the AHI stand in for the disease ignores the arousal index that frequently explains the patient's symptoms better. And in the laboratory, using a high-pass filter aggressive enough to attenuate frontal delta can erode the slow-wave activity on which N3 scoring depends, while too low a chin-EMG gain can hide the very RSWA that diagnoses RBD. In every case the corrective is the same: read the polysomnogram as an integrated physiological record, anchored in the EEG, and never let one channel or one number speak for the whole night.