The accurate diagnosis of events occurring during sleep requires knowledge of normal and abnormal sleep, skills in recording specialized forms of polysomnography, and experience in interpreting multiple physiological signals coupled to video and audio recordings. A careful history from the patient with a collateral history from a bed partner, parent, or caregiver is essential to determine whether a sleep study is necessary and to provide perspective into its subsequent significance. In some cases, the diagnosis may be obvious on history alone, whereas in others, a range of studies, including wake and sleep electroencephalography (EEG), brain magnetic resonance imaging (MRI), and neuropsychometric tests, may be needed. Solving the problems posed by complex parasomnias or sleep-related movement disorders may take considerable skills spanning the fields of sleep medicine, psychiatry, epileptology, and neurodegenerative disorders.

The recording technique most commonly used for elucidating nocturnal events is often referred to as video-EEG polysomnography, ¹ a term that reflects the multiple signals recorded during such studies. A standard polysomnogram (PSG) includes two electro-oculographic (EOG) derivations, three EEG derivations, channels recording chin and anterior tibial electromyography (EMG), surrogate measurements of airflow by nasal pressure and oronasal thermal sensors, measurement of oxyhemoglobin saturation by pulse or ear oximetry, assessment of thoracoabdominal movement by inductance plethysmography, and a recording of the electrocardiogram (EKG). For the recording of nocturnal spells, 16 EEG derivations are usually added to provide the ability to detect epileptiform abnormality. At least one additional EMG channel is included to record arm movements, often a derivation recording signals from the extensor digitorum communis muscles. To allow for adequate spatial resolution, only 16 channels at a time are usually displayed, but readers can move from the more conventional PSG display with an additional EMG channel to a display of whole-head EEG with relative ease using modern digital machines. Although EEG and EOG derivations are recorded with similar parameters to those used in a video-EEG monitoring (VEM) unit, the EMG pass-band is set at 10 to 100 Hz to reduce movement artifact without significantly affecting the amplitude of an interference pattern. Ideally, separate EMG derivations should be used to record right and left limb EMG, placing the electrodes 2 to 3 cm apart over the middle of the muscles; however, limitation on the number of channels often results in the use of one anterior tibial and one arm muscle derivation (often extensor digitorum), linking the two limbs. The main distinction from a conventional PSG, however, is the recording of video images and sound.

Although it is usual to monitor a patient undergoing a conventional sleep study by means of a camera and a microphone, the recordings are rarely stored or replayed during review of the study. When studying patients with nocturnal spells, however, the high-quality recording of behaviors is essential. Most digital PSG machines allow for real-time correlation of images and sounds with polygraphic tracings, permitting the exact temporal relationship between physiological signals and motor behaviors to be analyzed. This is important for a number of reasons. First, the PSG appearances of some of the arousal parasomnias, such as sleepwalking (somnambulism) and sleep terrors, are nonspecific, and the diagnosis can only be reached by seeing and hearing the event. Second, the presence of abnormally increased muscle tone during rapid eye movement (REM) sleep does not in itself make a diagnosis of REM sleep behavior disorder; abnormal motor activity must be seen if the history is inconclusive. Third, some parasomnias may be mimicked by arousals from sleep-disordered breathing, and the presence of crescendo snoring may be invaluable in determining the cause of an arousal. Fourth, other sounds during sleep may be of considerable diagnostic significance, such as those of stridor, bruxism, or the unusual parasomnia of catathrenia (sleep-related groaning).^2,3

Technologists recording nocturnal spells need special experience beyond that of an entry-level PSG or EEG technologist. Similarly, physicians supervising and interpreting such studies should be trained and experienced in scoring sleep stages, delineating obvious and subtle sleep-disordered breathing, and accurately identifying interictal and ictal EEG waveforms. Interpreters should feel comfortable reading studies at various playback speeds, ranging from 10-second screens for identifying epileptiform EEG to 30-second screens for sleep staging and 60- to 120-second screens for delineating sleep-disordered breathing. Attempts to diagnose parasomnias in sleep laboratories that mainly study obstructive sleep apnea or in EEG laboratories with little experience of sleep disorders usually fail, resulting in patients having to undergo repeated studies in a more specialized milieu.

Parasomnias are undesirable events or experiences that occur at sleep onset, during sleep, or during arousals from sleep.⁴ By convention, they exclude nocturnal seizures and movement disorders occurring during sleep, such as periodic limb movements, bruxism, and rhythmic movement disorder. They are classified as arousal disorders arising from non-REM (NREM) sleep, such as sleepwalking; parasomnias associated with REM sleep, such as REM sleep behavior disorder (RBD); and other parasomnias, such as sleep-related eating disorder (Table 22-1). During most parasomnias, complex nonstereotyped motor behavior occurs that may superficially resemble voluntary waking activity. Parasomnias most often studied by video-EEG polysomnography are discussed below.

Disorders of Arousal

The three disorders of arousal—sleepwalking, sleep terrors, and confusional arousals—all arise from NREM sleep, usually from the slow-wave (N3) stage. The conditions are postulated to result from a faulty arousal mechanism, leading to behaviors occurring in an abnormal state between slow-wave sleep and full wakefulness.^5,6 The exact pathophysiology is unknown, but a single-photon emission computed tomography (SPECT) study lends credence to the hypothesis that the events are occurring in a dissociative state with regional blood flow patterns differing from those found in slow-wave sleep or during wakefulness.⁷ During an episode of sleepwalking in this study, regional cerebral blood flow was increased in the anterior cerebellum and posterior cingulate cortex compared with slow-wave sleep, and blood flow was diminished in the frontoparietal association cortices compared with wakefulness. Predisposing factors include a family history of a similar disorder and age, with the condition most commonly manifesting in childhood. However, about 2 to 4 % of adults sleepwalk,^8,9 some persisting without remission from childhood, and some commencing de novo during adolescence or adulthood. Events on a particular night can be precipitated by factors increasing the probability of arousal, such as environmental noise, pyrexia, obstructive sleep apnea or stress; factors deepening slow-wave sleep, such as prior sleep deprivation, shift work or jet lag; and factors preventing full awakening, such as the use of central nervous system depressants.⁶

Sleepwalking is characterized by complex, coordinated motor behavior resulting in the patient’s arising from the bed and walking. Consciousness is impaired, and the sleepwalker may be hard to arouse with little or no recollection of the events later. Behaviors are frequently inappropriate, such as urination in a closet or climbing into someone else’s bed. Although most sleepwalking does not result in injuries, violent behaviors may occasionally occur,¹⁰ including assaults on others, self-injury by falling down stairs, or jumping through a window.

Sleep terrors consist of arousals from slow-wave sleep accompanied by motor, vocal, and autonomic features of intense fear. Tachycardia, diaphoresis, and a piercing scream usually characterize the events. The patient is usually unresponsive and has little recall of the event. Either no preceding dream can be remembered, or the patient has a foggy image of a frightening or dangerous dreamt situation, such as the bed being on fire.⁴

Confusional arousals consist of disorientation on awakening from slow-wave sleep with neither the complex motility of sleepwalking nor the intense fear of sleep terrors.⁴ The patient will often sit up with a bewildered expression, talk unintelligibly, and appear unresponsive before returning to sleep.

The diagnosis of an arousal parasomnia is usually made on history without the need for sleep studies. However, under certain circumstances, video-EEG PSG is required. These include a history of violent, potentially injurious, or disruptive behaviors requiring precision of diagnosis, or uncertainty whether the behaviors are due to a disorder of arousal as opposed to seizures or REM sleep behavior disorder.¹¹ The primary goal of the sleep study is to record a typical event. However, episodes do not occur regularly and often not during a single night study. To precipitate events, laboratory protocols, such as producing a loud noise during slow-wave sleep, are employed to cause a sudden arousal.¹² Another approach is to perform the PSG after 25 or more hours of sleep deprivation.¹³ Slow-wave sleep increases in duration at the expense of all other stages of sleep, as does the number of arousals from slow-wave sleep. A greater percentage of these arousals are associated with somnambulism, with one study reporting 90% of patients demonstrating a behavioral event during recovery sleep, compared with only 50% during the baseline night. Control subjects without a history of arousal disorder do not manifest sleepwalking during the recovery night.¹⁴

When a typical event fails to occur, readers must review the entire PSG, especially at the time of any arousals from NREM sleep. The videotape may demonstrate minor confusional arousals that may be missed by even experienced technologists. The video recording will show a sudden arousal, with the patient partially or completely sitting with a confused expression. Periods of REM sleep should be carefully reviewed to assess muscle tone, which, if normal, usually excludes a diagnosis of REM sleep behavior disorder. The additional 16 EEG derivations should be read at a speed of 10 seconds per screen to determine whether potentially epileptogenic spikes or sharp waves are present, which might provide supportive evidence for a seizure disorder. Background sleep architecture in patients with disorders of arousal may statistically differ from normal architecture, with higher percentage slow-wave sleep, more even distribution of slow-wave sleep through the night, and higher delta power in the first sleep cycle.^15–17 However, individual variability precludes these phenomena from being diagnostically helpful.

The PSG appearance of the three different types of disorders of arousal is largely identical (Figure 22-1). A period of slow-wave sleep ends abruptly with a burst of generalized muscle activity. In 47% of events, the arousal is preceded by a single so-called hypersynchronous delta wave in the central EEG derivation.¹⁸ This is a high-amplitude, more rhythmic delta wave standing out from the background slow-wave activity. Two or more hypersynchronous delta waves occur in only 16% of episodes, and delta waves are rarely present in more than one derivation (2%). These waves are nonepileptiform, and most probably represent nonspecific arousal complexes. Their absence does not dissuade from the diagnosis. Respiratory channels should be examined to be certain that the parasomnia is not precipitated by upper airway obstruction. Following the arousal, tachycardia may develop. A study of 252 arousals from slow-wave sleep showed that the EEG in the first 10 seconds after arousal demonstrated one of several patterns: predominant alpha and beta frequencies (36%), persistent delta waves (6%), mixed delta and theta frequencies (38%), or muscle artifact obliterating the EEG waveforms (20%).¹⁸ Similar findings were seen in a later independent study.¹⁹ In particular, these studies demonstrate that events occurring during alpha rhythm may be due to an organic disorder of arousal and should not be automatically considered psychogenic in origin.