Although often difficult to obtain, a precise history of events can be very helpful in distinguishing disorders of arousal from seizures.

Patients presenting with abnormal nocturnal behaviors often pose a diagnostic challenge. Detailed history is sometimes difficult to obtain, because patients with parasomnia and those with nocturnal seizures may be amnestic for the event and often there are no witnesses to provide important details. Electroencephalography (EEG) can be helpful, as the abnormalities seen in epilepsy are highly specific.


Per ICSD II definition,1 “parasomnias are undesirable physical events or experiences that occur during entry to sleep, within sleep, or during arousals from sleep.” Classified by sleep stage when the abnormal behaviors occur, they may be NREM parasomnia/disorders of arousal, parasomnias associated with REM sleep, or other.

Nocturnal seizures may also present abnormal nocturnal behaviors. These tend to be stereotypic and may occur with abrupt awakening from sleep. Presentation may include generalized tonic-clonic movements, urinary incontinence or tongue biting, automatisms, focal/stereotypic limb movements, or facial twitching; the seizure may be followed by postictal confusion or transient paralysis. Typically, ictal EEG abnormalities are seen during the event. Between events, there are often interictal EEG abnormalities, and these are highly specific for epilepsy patients.

Distinguishing nocturnal seizures from parasomnia has important therapeutic implications. EEG abnormalities can help diagnose epilepsy, but there are no known EEG features specific for disorders of arousal.


Initial history
Mr DG was a healthy 36-year-old man who presented for consultation at our sleep clinic after injuring his arm during sleep. He described that he had had abnormal nocturnal behaviors occurring intermittently for more than a decade, but did not seek medical attention for them until the arm injury. His memory for any given behavior was very faint. Since he did not have a stable bed partner, the initial history provided was relatively sparse. He thought most events would occur in the first third of the night. At the time of the arm injury, he had been seen in the emergency department at 3 am, and hypothesized that the behavior probably had occurred at about 1 to 2 am.

He reported that the nocturnal behaviors were variable—such as punching or hitting objects, jumping out of bed, or walking to a different part of the room. On one occasion, he may have gone outside the room. In one incident, he hit his girlfriend, which prompted an awakening. He did not recall a dream. He and his girlfriend discussed the situation, and both eventually laughed at the incident and promptly fell asleep afterwards.

The behaviors occurred at unpredictable intervals and were sometimes precipitated by stress. He was less likely to have an event when he spent time in his summer house or when he was with his girlfriend.

He occasionally had fragmented sleep, but denied other major sleep complaints, such as periodic motor activity at night, difficulty with sleep initiation, significant snoring, sleep paralysis, or cataplexy. His typical bedtime was around midnight, and wake time varied between 7 and 9 am.

The patient had a normal birth and development, no history of major head traumas, and no central nervous system illnesses; he never had a febrile seizure or any other convulsion, or any episodes of altered awareness. There was no history of mental illness.

His other medical problems included a recently diagnosed mild hypothyroidism and acne. He was otherwise healthy.

His medications included levothyroxine, minocycline, and topical creams. Levothyroxine was started recently, and dose was being adjusted at the time of the event that led to the injury.

The patient’s family history was significant for sleepwalking in his father. Notably, sleepwalking both in the patient and in his father started in adulthood.

The patient was a nonsmoker and did not use alcohol or caffeine in excess or any recreational substances. He was working full time at his own marketing company.

Physical exam
The patient was 71 inches tall and weighed 221 pounds. His blood pressure was 129/78 and heart rate was 94. Oxygen saturation was 97% on room air. Neck circumference was 16¾ inches. He was a pleasant, cooperative man in no acute distress. His upper airway was mildly narrowed to a Mallampati type III. The rest of the HEENT (heart, eyes, ears, neck, and throat examination) was unremarkable. Chest was clear to auscultation. Heart auscultation showed no murmurs, rubs, or gallops. His mental status was intact: language was fluent without paraphasic errors, and he had normal spatial orientation and memory for recent events. Cranial nerves were II through XII within normal limits. His strength and sensation were within normal limits, and reflexes were symmetric without focal pathology. His gait and coordination were normal.

Figure 1. Brain MRI evaluation, prompted by the concerning EEG findings, was normal.


The patient underwent an attended polysomnographic recording of sleep pattern (EEG, EOG, EMG), airflow (oronasal thermistor and nasal pressure), snoring, respiratory effort (thoracic and abdominal motion), arterial oxygen saturation, body position, ECG, and leg movements (anterior tibialis EMG). Video-EEG monitoring was performed throughout the recording using anterior temporal as well as standard 10-20 EEG electrodes. Spike and seizure detection was performed visually and with the aid of a computer program to rule out any epileptiform features.

The patient had an initial sleep latency of 15 minutes and a sleep efficiency of 86.6%. Sleep architecture was unremarkable for age, with a slightly increased proportion of stage 1 and 2 sleep (70.9%), a normal proportion of slow wave sleep (12.8%), and a slightly decreased proportion of rapid eye movement (REM) sleep (16.3%). REM sleep latency was decreased at 63 minutes. REM atonia was well preserved.

There were no abnormal limb movements and no major ECG abnormalities noted.

Respiratory evaluation
Mild obstructive sleep apnea (OSA) was observed, with an apnea-hypopnea index of 10.8 events/hour. The respiratory events produced sleep fragmentation and only mild oxyhemoglobin desaturation. Intermittent soft snoring was recorded.

Video-EEG monitoring
The waking EEG background was well organized, consisting of well-formed and modulated 30-50 µV, 10 Hz alpha, and moderate bilateral beta. During sleep, normal sleep complexes were seen. There were no electrographic seizures and no abnormal behaviors.

There was one abnormally shaped K-complex with imbedded left temporal sharp wave (Figure 1).

At a follow-up visit, the patient brought his girlfriend to aid with the description of the nocturnal events. She reported a highly variable pattern of behavior. She had observed instances in which the patient would remain in bed and have abrupt, brief, punching or hitting type movements of the arms. On other occasions, he would suddenly jump out of bed and turn on the lights.

On a third occasion, the patient had walked to the window and tried to open the shades. When she awakened and questioned him, he reported a fragmentary sensation that he must have dreamed about needing additional light, but could not recall any elaborate dream associated with the behavior.

On some occasions, the nocturnal behaviors involved sleep-talking only—with variable phrases, words, and content used.

The duration of each event was variable, lasting between several minutes and about 10 to 15 minutes. When fully awakened from an episode, the patient acted appropriately with no impairment of speech and no focal neurological symptoms.

During a month of relatively increased stress (due to relocation), both noted a total of five events, each one with different presentation. The events occurred only once per night.


Our working diagnosis was a disorder of arousal—NREM parasomnia, variant of sleepwalking. Differential diagnosis included complex partial seizures and postictal confusion episodes—considered less likely after the more detailed history of the events was reported, due to the clinical description of the events, their duration, variability, lack of frank convulsive episodes, and lack of significant seizure risk factors. REM behavior disorder (RBD) was not likely, given the presentation and the lack of dream association, as well as the well-preserved atonia of REM sleep seen on the polysomnogram.


After we discussed the findings from the evaluation in detail, as well as factors that commonly lead to sleep fragmentation, a decision was made to start pharmacologic treatment.

Given the concerning EEG findings, we aimed at using agents that would not lower seizure threshold and, if possible, have antiepileptic properties. Clonazepam was started, but led to excessive sedation and therefore was replaced with lorazepam. The initial dose of 0.5 mg was not sufficiently effective, but at 1 mg per night the frequency and intensity of events were markedly reduced.

The mild sleep apnea was treated conservatively with positional therapy, avoidance of alcohol, weight control, and later an oral appliance.


Eighteen months after treatment onset, only rare events occurred—mostly provoked by stress or sleep loss. He reported that sleep is more consolidated—after starting lorazepam and further improved after treatment of his mild OSA.


This patient had a history highly suggestive of sleepwalking, yet the abnormalities on EEG were very concerning for epilepsy. How do we distinguish the two conditions and optimize treatment?

Although quite rare, epileptiform abnormalities can be seen on EEG even in individuals who have never had a seizure. For example, in a study of 970 healthy children, Capdevila et al2 found that the prevalence of epileptiform activity was 1.45%. In another study of 1,057 children aged 6 to 12 years, which also included hyperventilation during the EEG, Okubo et al3 found that 5% of children had epileptiform discharges. In a study of adults, Gregory et al4 found that only 0.5% had unequivocal epileptiform abnormalities. However, in this later study, abnormalities with uncertain significance were excluded (and possibly our patient may have been among them). Thus, epileptiform EEG abnormalities are highly specific for patients with epilepsy.

It is possible that our patient is among the few (1%-5%) individuals, who may have an EEG abnormality and no seizures. There was only one abnormality seen throughout the overnight recording of more than 6 hours. A typical routine EEG recording is less than 1 hour long, and therefore patients with rare abnormalities (such as our patient) may not have been included in the statistics cited above. Additionally, the shape of the abnormality, although concerning because it is sharp and asymmetric, has a broad central focus, and this makes it more difficult to interpret physiologically. Therefore, it should be taken in the context of the specific clinical presentation.

On the other hand, epilepsy and sleep disorders frequently coexist in the same patient. Specifically, Silvestri et al5 described six patients with a history of sleepwalking, who subsequently developed epilepsy. Another case report6 described a patient with a 14-year history of epilepsy who subsequently developed parasomnia. Thus, we should be cautious and allow thorough evaluation and close follow-up of our patients.

Although often difficult to obtain, a precise history of the events can be very helpful in distinguishing disorders of arousal from seizures. In all the above reports, the presentation of the parasomnia is semiologically different from the epileptiform events. Additionally, Zucconi et al7 described the presentation of nocturnal motor phenomena in patients with and without epilepsy, and in this study the motor phenomena in epilepsy patients were different than those in the patients without epilepsy. Thus, semiology is crucial in distinguishing the disorders of arousal from nocturnal epileptic seizures.

Parasomnia is particularly difficult to distinguish from frontal lobe epilepsy, as in this form of epilepsy, the interictal EEG abnormalities are rare and may not be seen even on a full night recording. Manni et al8 have tested a scale comprised of typical clinical criteria for seizures, such as stereotypic presentation, duration of events, presence of dystonic posture, confusion, and other characteristics. A detailed copy of the scale they used is provided at the end of the report. In a study of 71 subjects with epilepsy, RBD, or arousal parasomnia, they found the scale to have a sensitivity as a diagnostic test for nocturnal frontal lobe epilepsy (NFLE) of 71.4%; the specificity was 100%; the positive predictive value was 100%; and the negative predictive value was 91.1%. If graded on this scale, our patient would have a very low score of -5, making the likelihood of nocturnal epilepsy very low on clinical grounds.

It is notable that in this case the abnormality occurred only once in the whole recording and it was imbedded in a K-complex. K complexes can be triggered by auditory stimulation and often precede arousals. Thus, we could speculate that an abnormal arousal mechanism may have contributed to the EEG abnormality we saw in this K-complex. However, such identifying features have not been previously reported among patients with parasomnia.

In summary, the diagnosis in this case was based on the clinical presentation, taken in context with the polysomnography findings. A detailed history of the events was key in guiding the diagnosis and treatment. The presence of EEG abnormalities was concerning, and thus we performed additional clinical and MRI evaluations and closer follow-up. Further systematic investigations of the EEG of patients with disorders of arousal and without epilepsy may help establish specific EEG findings that identify these patients.

Milena Pavlova, MD, is an instructor in neurology at the Brigham and Women’s Hospital/Harvard Medical School, Boston. She is also the medical director at the Faulkner Hospital Sleep Center. She is involved in multiple research projects on the interaction between sleep and medical and neurological disorders, including sleep, circadian rhythms, and epilepsy. Recent publications include “Is There a Circadian Variation of Epileptiform Abnormalities in Idiopathic Generalized Epilepsy?” (Epilepsy and Behavior. 2009;16:461-7). She can be reached at


  1. American Academy of Sleep Medicine. International Classification of Sleep Disorders: Diagnostic and Coding Manual. 2nd ed. Westchester, Ill: AASM; 2005
  2. Capdevila OS, Dayyat E, Kheirandish-Gozal L, Gozal D. Prevalence of epileptiform activity in healthy children during sleep [published online ahead of print July 16, 2007]. Sleep Med. 2008;9(3):303-9.
  3. Okubo Y, Matsuura M, Asai T, et al. Epileptiform EEG discharges in healthy children: prevalence, emotional and behavioral correlates, and genetic influences. Epilepsia. 1994;35(4):832-41.
  4. Gregory RP, Oates T, Merry RT. Electroencephalogram epileptiform abnormalities in candidates for aircrew training. Electroencephalogr Clin Neurophysiol. 1993;86(1):75-7.
  5. Silvestri R, de Domenico P, Mento G, Laganà A, di Perri R. Epileptic seizures in subjects previously affected by disorders of arousal. Neurophysiol Clin. 1995;25(1):19-27.
  6. Liebman RF, Rodriguez AJ. A patient with epilepsy and new onset of nocturnal symptoms. Rev Neurol Dis. 2009;6(1):37-8.
  7. Zucconi M, Oldani A, Ferini-Strambi L, Bizzozero D, Smirne S. Nocturnal paroxysmal arousals with motor behaviors during sleep: frontal lobe epilepsy or parasomnia? J Clin Neurophysiol. 1997;14(6):513-22.
  8. Manni R, Terzaghi M, Repetto A. The FLEP scale in diagnosing nocturnal frontal lobe epilepsy, NREM and REM parasomnias: data from a tertiary sleep and epilepsy unit [published online ahead of print April 10, 2008]. Epilepsia. 2008;49(9):1581-5.