An anoxic seizure is a consequence of a syncope, which is an abrupt cutting off of the energy substrates to the cerebral cortex, usually through a sudden decrease in cerebral perfusion by oxygenated blood (1,3). The term anoxic seizure is shorthand for the clinical or electroclinical event that occurs as a result of the cessation of nutrition to the most metabolically active neurons. Less complete, or less rapidly evolving syncope will have less dramatic consequences. For ease of reference, the syncopes have been subdivided, but it must be recognized that overlaps occur and present knowledge is inevitably incomplete. In particular, it seems likely that what we call reflex anoxic seizures or reflex asystolic syncope (RAS), breath-holding spells or prolonged expiratory apnea, vasovagal syncope, and neurocardiogenic syncope are all varieties of the same disorder, called by the adult cardiologists neurally mediated syncope. When parents or children are bewildered by these diagnoses, or annoyed when told that it's only breath holding or only a simple faint, then contact with a family support organization may be very helpful (http://www.stars.org.uk).

Gastaut (10) used the term reflex anoxic cerebral seizures to describe all the various syncopes, sobbing spasms, and breath-holding spells that followed noxious stimuli in young children. Since 1978, reflex anoxic seizure has been used more specifically to describe a particular type of nonepileptic convulsive event, most commonly induced in young children by an unexpected bump to the head (11). Although other terminology, such as pallid breath-holding and pallid infantile syncope, have been applied to such episodes (12), the term reflex anoxic seizure is now widely recognized (13,14).

Until the advent of cardiac loop recorders, direct evidence with respect to the pathophysiology of natural attacks had been very limited. Since prolonged cardiac recording has become feasible in children, many recordings of prolonged reflex asystole have been recorded and several examples published (1518).

An extract from a letter written to a pediatrician by a consultant neurologist may give the reader some idea of the diagnostic difficulties that were commonly experienced before the phenomenon of reflex asystole was well known. He wrote:

Thank you for asking me to see this 7-year-old young man. As a toddler he began to have attacks of loss of awareness, rigidity, and eye rolling which would be induced by minor knocks. This has continued and recently an episode occurred in which he had an undoubted tonic-clonic seizure with incontinence of urine. Curiously, as far as I can tell from mother's account, every attack has been triggered by a minor bump on the head, and he has never had an attack out of the blue. He had difficulties at birth. The family history is clear except for a convulsion in the mother when she was tiny, about which there is no further information. It seems to me that this boy is having a form of reflex epileptic seizure, and my inclination would have been to start treatment with sodium valproate. In fact mother told me . . . that he was started on Epilim just a couple of weeks ago. Even though two EEGs have been normal, I do not doubt that he has an epileptic tendency, and I am sure that he should be on treatment for at least a couple of years free from attacks.

When this boy was seen in a further consultation he was "an epileptic," his school knew about his "epilepsy," his mother was in touch with an epilepsy association, and invalidity benefit had been applied for on the basis of epilepsy. Presumably, the difficulty here was that neither the pediatrician nor the neurologist knew that this was precisely the story of nonepileptic reflex anoxic seizures of vagally mediated cardioinhibitory type, otherwise known as reflex asystolic syncope. It is probable that the diagnosis of breath-holding spells had been entertained earlier, but quite rightly discarded, if only because the boy was by now over the age of 7 years. The alternative diagnosis of a primary cardiac syncope, such as is seen in the long QT syndrome (see Long QT Disorders below)LQTwas not considered, perhaps because syncopes in LQT had not at that time been described exclusively as a sequel to minor bumps to the head. Since then, there has been a report of one case of LQT in which four or five syncopes, the last fatal, were precipitated by a blow to the head (19: page 94).

As children grow older, reflex anoxic seizures may cease altogether or change to more obvious convulsive or nonconvulsive vasovagal syncope in childhood and adolescence. It is possible, although proper long-term studies have not been done, that syncopes may reappear in old age.

Beyond the toddler stage, children may report sensory disturbances along with the syncopes. Most dramatic are out-of-body experiences with a dreamlike quality (17), which may include the child feeling as if he or she has floated up to the ceiling and is watching his or her body lying on the floor in a seizure (20). Night terrors (see Parasomnias below) as a sequel to the syncopal episodes have also been reported by parents.

It is often said that breath-holding spells (see Breath-holding Attacks below) or syncope caused by prolonged expiratory apnea may also occur in children who have reflex anoxic seizures. It is certainly true that in some children some episodes may be more blue or cyanotic and some more pale and blanched looking, but there are no good recordings that confirm this proposition.

It is best to try to make a precise diagnosis as to whether a convulsive syncope in a young child is cardiogenic or respiratory in origin. If it is cardiogenic, then the main differential diagnosis is a reflex anoxic seizure (reflex asystolic syncope) versus a convulsive syncope from long QT syndrome or other cardiac cause. If it appears to be a respiratory (i.e. apneic) syncope, then the differential diagnosis is breath-holding spells (prolonged end-expiratory apnea) or suffocation (in particular, from imposed upper-airway obstruction as part of Meadow's syndrome (see Suffocation below).

Vasovagal syncope is the most familiar and predominant form of neurally mediated syncope. If classical reflex anoxic seizures (with reflex asystolic syncope) represent a fairly pure vagal attack, vasovagal syncope involves a vasodepressor component with variable vagal accompaniment. Episodes may begin in infancy, sometimes with reflex anoxic seizures, and thereafter are seen at all ages, becoming most dramatic perhaps in old age (21).

Tables in medical textbooks or works of epileptology tend to perpetuate gross errors with respect to the distinction between vasovagal syncope and epileptic seizures with comparable signs. This may be in part because many authors equate syncope with some sort of limp, pallid swooning in the Victorian manner. Here, for example, are the features previously said to distinguish between syncope and seizures:

• Posture: upright
• Pallor and sweating: invariable
• Onset: gradual
• Injury: rare
• Convulsive jerks: rare
• Incontinence: rare
• Unconsciousness: seconds
• Recovery: rapid
• Postictal confusion: rare
• Frequency: infrequent
• Precipitating factors: crowded places, lack of food, and unpleasant circumstances.

In reality the situation is different. Vasovagal syncope may occur supine, particularly in the case of venepuncture fits; though some (13) would call these reflex anoxic seizures insofar as the mechanism is strongly cardioinhibitory. Pallor and sweating are certainly not invariable, nor need onset be gradual. There is no difference between the liability to injury in convulsive syncope as opposed to a comparable convulsive epileptic seizure. Convulsive jerks are certainly not rare but occur in perhaps 50% of vasovagal syncopes (22) and more often in experimental syncope (23). Urinary incontinence is common (24), occurring in 10% of cases in one experimental study (1). Unconsciousness may be much more than seconds and recovery, although it may be rapid in mild syncope, is not necessarily complete early on (3). It is true that postictal confusion proper is rare, but it can occur (3). The frequency of vasovagal syncope may be very great, up to more than once a day. Stimuli may be very subtle, but it is true that some sort of stimulus should be detected for at least some attacks in any individual.

The setting and stimulus are indeed the most important factors in allowing the presumptive diagnosis of vasovagal syncope, together with elicitation of the warning symptoms or aura, which are commonly present. A seizure that occurs after a bath while the child is having her hair blow-dried, or brushed, iswithout need for further investigationa vasovagal nonepileptic convulsive syncope. Premonitory symptoms are usually present in older children, even if the duration is only a second or two, but sometimes these are forgotten and only recalled when syncope is reproduced, as in the head-up tilt test. All physicians are aware of the usual symptoms of cerebral ischemia, such as dizziness and greying out of vision and tinnitus, but an important additional symptom is abdominal pain. It may be difficult to tell whether abdominal pain is a symptom or trigger of a vasovagal syncope or an intestinal symptom of a strong vagal discharge. The latter is quite common (1), and sometimes leads to confusion with the so-called epigastic aura, which may precede the complex partial component of the temporal lobe epileptic seizure. Almost all children with vasovagal syncope have a first-degree relative, commonly a parent, affected (25). It is unfortunately common [histories are included in (1)] to find that the parent who now seems convincingly to have vasovagal convulsive syncope has become irredeemably "epileptic" and too habituated (or too frightened of losing a precious driving license) to discontinue years of useless (and perhaps embryopathic) antiepileptic medication.

Such considerations have led to the use of head-up tilt testing, not only as a diagnostic aid but as a diagnostic reinforcer. If, for example, a child patient, a family doctor, and a pediatrician have all been convinced that the child has had "grand mal epilepsy" since the age of 3, then some dramatic theater may be necessary at the age of 12 to 14 years to make the switch from epilepsy to the correct diagnosis of vasovagal syncope. There are no good data on this point, but there is an impression that if the diagnosis is not properly instilled by this age, it may be too late to prevent a life of being "epileptic".

A case history illustrates the transition from reflex anoxic seizures in infancy through short latency pain-induced vasovagal syncope to blood-injury phobia (26,27) in adolescence: the history was given by the mother when her affected daughter was aged 13 years. The previous diagnoses had included epilepsy, hypoglycemia, and hysterical behavior.

The first episode occurred at the age of 10 months, after a very slight bump to the infant's head. The appearance of the attacks has been similar from then to now, except that severity has varied and tended to increase with the passage of time. Typically there is a latency of 10 to 20 seconds during which she may say "oh mum I've hurt myself." By this time the blood has drained from her face, she goes limp, and falls as if dead, then goes totally rigid, making a noise like a cackle or gurgle, with her hands and feet turned in and her back sometimes forming the shape of an arc. Sometimes her arms and legs jerk, but not violently, as though pedaling her bicycle, but on occasion thrashing wildly like a full seizure (as her mother describes it). Again, she looks like death and then wakes up as if coming out of a very deep sleep. She is then very disorientated, does not know what has happened or where she is, but within a couple of minutes she has come to herself and may then want to lie down again and have a proper sleep. Since about the age of 7 or 8 years she has described an aura. She hears a noise like a high-pitched screaming and sometimes hears a voice but cannot describe the voice precisely. Sometimes she sees red, a color she does not like. More recently she has had strange hallucinations during the warning period, such as seeing a train rushing towards her. The stimuli have modified over the years after the first head bump. All episodes in earlier years followed small pains like her finger being bent back. Then she developed the same reaction to seeing a minor injury such as a scab that had come off a wound, and then inevitable syncope at the sight of blood. Most recently merely the thought of self-injury was sufficient.

On the evening before the intended consultation she was told (wrongly) that her eyeballs would be pressed down (ocular compression) and within 2 minutes she was stiff and snorting. Although the mother's sister had had some type of genuine epilepsy, a family history of syncope of any kind was denied. Actually, the mother later admitted to several faints during adolescence and pregnancy, but did not mention them because she did not have a "fit." The results of a recent study suggest that adults with blood or injury phobia have a "constitutional autonomic dysregulation" predisposing them to neurally mediated syncope even in the absence of any blood or injury stimulus, and that repeated syncopes resulting from such stimuli secondarily lead to the blood or injury phobia (28).

By contrast to vasovagal syncope, convincing vago-vagal syncope is rare. The reflex is usually triggered by swallowing or vomiting and cardiac standstill results, with a motor anoxic seizure (convulsive syncope) if the asystole is sufficiently prolonged. This is probably not a life-threatening disorder, but the symptoms can be troublesome, particularly if the patient also has migraine with associated vomiting. Pacemaker therapy has been used successfully in this situation (1).

Hyperventilation in any human induces various organic symptoms that may in certain individuals stimulate further hyperventilation and exacerbation of the original symptoms. A degree of panic may be so engendered. Asking the child to hyperventilate (whether by getting the child to repeatedly blow out a candle, blow soap bubbles, blow a tissue, or to directly hyperventilate) may induce symptoms similar to those of which the child complains. Continuation of hyperventilation once the directed hyperventilation has been stopped may be of additional diagnostic value. Spontaneous hyperventilation may lead to apparent absences without spike and wave (29), but it should be remembered that a possible difficult-to-diagnose absence-like seizure may be of frontal lobe origin (30). Studies on hypocapnea and the cerebral circulation include those found in references 31 and 32.

Syncope due to orthostatic hypotension secondary to autonomic failure is rare in childhood. Dopamine b-decarboxylase deficiency (33) is a possibility in such a clinical situation. The simplest way of detecting orthostatic intolerance is to stand the child on a foam mat (to avoid injury when falling) for 10 minutes with continuous blood pressure measurements; this is best done using Finapres recording from a finger with the hand secured at heart level. This method may also be used to provoke vasovagal syncope (see Vasovagal Syncope above) in young children, including those too young to tilt (34).

Chronic orthostatic intolerance can produce other symptoms, usually in addition to vasovagal syncope, these include symptoms of presyncope: lightheadedness, "dizziness," blurred vision. Furthermore exercise intolerance, chronic fatigue, migrainous headache, nausea, abdominal discomfort, chest discomfort, palpitations, shortness of breath, hyperventilation, peripheral cyanosis, sweating, and flushing on standing have been encompassed in this condition (35). Clues may be tiredness and excessive dislike of exercise. Chronic orthostatic intolerance is sometimes part of the clinical picture in chronic fatigue syndromes and it maybe helpful to consider this treatable disorder as a differential of idiopathic chronic fatigue syndrome.

One clinical picture comprising chronic orthostatic intolerance in teenagers and young adults is the postural orthostatic tachycardia syndrome (POTS) (36). Patients have symptoms of chronic orthostatic intolerance with significant daily disability, associated with a marked tachycardia on standing: a heart rate increase of >30 beats per minute or a heart rate of >120 beats per minute within 10 minutes of head-up tilt (35).

The long QT syndromes are associated with genuinely life-threatening syncopes that may be hypotonic or convulsive. The mechanism of the syncopes is a ventricular tachyarrhythmia, normally torsades de pointes. As a rule, there is no great difficulty in the diagnosis of the syndrome of Jervell and Lange-Nielsen (37), in which congenital deafness is associated with an autosomal recessive inheritance. Much more difficult is the Romano-Ward syndrome (38), which is dominantly inherited but with incomplete penetrance. It has been suggested (39) that the diagnosis may fairly easily be made by asking the right question, in particular asking whether the child lost consciousness and remained completely still (like a dead body) for several seconds before having "tonic-clonic seizures." Actually what these authors describe is a cardiogenic syncope, not fundamentally different from the reflex anoxic seizure seen with reflex asystolic syncope or a convulsive vasovagal syncope in which the vagal component predominates. The observation that the child lies like a dead body is not necessarily made, and of course the nonepileptic seizure is normally not a tonic-clonic seizure but an anoxic seizure with a combination of spasms and jerks and stiffening. A degree of overlap exists between the stimuli that induce the neurally mediated syncopes and those that trigger the ventricular tachyarrhythmias of the long QT syndrome, but the most important hint in favor of a long QT disorder is the story of convulsions triggered by fear or fright and particularly in two situations:

• During exercise, especially when that exercise is emotionally charged
• During sleep

A personal example illustrates diagnostic difficulties:

A 5-year-old girl presented with a history of convulsive syncope since the age of 2. At the first consultation the parents said that when she fell, not necessarily hurting herself and not necessarily falling on any particular part of her, she went gray or gray/purple around the mouth, looked faintish as if dead, went very rigid as her eyes rolled and her head flopped, she moaned, and "she was dead in my arms." One of the episodes was said to have occurred as a splinter was being taken out of her finger by her mother. There was a positive family history in that the father had fainted on cutting his finger and the mother had faints during pregnancy. An interictal 24-hour ambulatory cassette ECG of the child had been reported as normal. A diagnosis of reflex anoxic seizures (reflex asystolic syncope) was made and it was decided that it would not be necessary to do ocular compression as a confirmatory test. Three years later the consultant pediatrician wrote again:

She had approximately 1 year without any episodes, but has had two close episodes in the last few weeks, both of which occurred during physical exertion during play. At least one of these episodes seemed to be associated with an olfactory aura, the child describing strange smells before the event. In both situations, she was found unconscious, stiff, and mottled gray but recovered fairly promptly. I guess this is still a vagally mediated event, but the parents would value further assessment and reassurance.

Review of the history revealed that although two of the episodes had originally been associated with falling when playing with a ball, other episodes had occurred when chasing a dog, trying to catch the waves at the edge of the sea, playing being chased on her bicycle, and during a hopping race. The new historical details prompted immediate measurement of her QT interval, the corrected value of which (QTc) was 479 milliseconds (normal value less than 440 milliseconds). A review of the original 24-hour ECG from 3 years previously showed that the QTc was prolonged then also at 470 milliseconds. Her mother had a marginally prolonged QTc of 449 milliseconds, whereas her father and sister had normal QTc measurements of 387 and 390 milliseconds respectively.

Long QT disorders are much less common than neurally mediated syncopes, such as reflex anoxic seizures and reflex asystolic syncope, but this diagnosis should be sought when the precipitants of a cardiac syncope are not of the typical benign reflex anoxic seizure type (that is to say, unexpected bumps to the head) and particularly when exercise or sleep are triggers. This diagnostic consideration is another reason for trying to separate by history cardiac and respiratory syncopes.

Having said this, provided that the history is typical of a variety of neurally mediated syncope and a careful cardiac evaluation is negative, one should be reassuring even when syncopes occur during exercise (40).

Diagnostic difficulties do not usually arise with respect to endogenous cardiac syncopes other than those of the long QT syndromes. However, it is up to the clinician to obtain a sufficiently clear history to determine whether a seizure or convulsion is an epileptic seizure or is a nonepileptic convulsive syncope. Sometimes ventricular tachyarhythmias occur with normal QT intervals (4144), and there are occasions in obvious congenital heart disease when, for example, paroxysmal pulmonary hypertension may have to be inferred by a precise description, indicating an anoxic seizure precipitated by exercise (1).

Breath-holding spells have been described for centuries (45) but controversy as to what they are remains (46). The term "breath-holding" is not at all satisfactory (19,47). It tends to give offence to parents of affected children. It seems to imply temper tantrums and bad behavior. One imagines that many members of the public and even pediatricians actually do believe that breath-holding spells are a manifestation of a behavior disorder and, in some pediatric textbooks, breath-holding attacks are to be found in the section on psychiatric or psychological disorders. However, studies have shown that however one defines breath-holding spells, behavioral disorders in those afflicted do not differ from those in control children (48).

Breath-holding seems to imply some sort of voluntary "I'll hold my breath until I get what I want" behavior, whereas none of the behaviors so described seems to involve this mechanism. There is no difficulty nowadays in recognizing that what used to be called white or pallid breath-holding (12) has a cardiac rather than a respiratory mechanism, as discussed earlier in the section on reflex anoxic seizures and reflex asystolic syncope. The term prolonged expiratory apnea (49) is certainly helpful in discussing those episodes in which the mechanism is predominantly respiratory, even though the pathophysiological details may be in dispute.

One difficulty is, as with so many paroxysmal disorders, that precise detailed documentation of what happens is in short supply. Cinematographic registration has been described (50). Videorecordingspredominantly of several episodes in a single childhave been obtained (1,51) and polygraphic recordings of a few children (52), but the total information compared to the frequency of occurrence of natural episodes is very small. There appears to be a pure respiratory "breath-holding" spell or prolonged expiratory apnea, without any change in cardiac rate or rhythm (albeit information on cardiac output is not available), such attacks being clearly cyanotic or "blue" breath-holding. There are also episodes that may be described as "mixed" breath-holding, insofar as there is not only expiratory apnea but also a degree of bradycardia or cardiac asystole (1,51).

An argument exists about the prognosis of these "cyanotic breath-holding spells" or prolonged expiratory apneas (51,52), but management of neurodevelopmentally intact children does depend on the general assumption that prolonged expiratory apnea (cyanotic breath-holding) is benign (46). The best prospective study to date is that of DiMario (53), albeit he includes pallid breath-holding spells, which may represent, in our terminology, reflex anoxic seizures or reflex asystolic syncope.

Children with aberrant development, including those with autistic disorders, may have atonic or more dramatic syncopal seizures compulsively self-induced by something akin to a Valsalva or Weber maneuver (3,54). Such episodes may be very severe and, indeed, may have a fatal outcome (55). The child seems able to obstruct the cerebral circulation completely, so that an anoxic seizure results. Perhaps, because of the cerebral abnormality already present, this is one situation in which anoxic-epileptic seizures (see Anoxic-Epileptic Seizures below) may result (5658). If the episodes are very frequent, as is often the case, detailed analysis by videorecording and polygraphic registration (55) may allow a precise elucidation of the diagnosis. Clues include the video-picture of true "breath-holding" for about 10 seconds, this time in inspiration, reduction of the amplitude of the QRS complexes on ECG, and then a burst of high-voltage slow waves on EEG. Sometimes, hyperventilation precedes the Valsalva maneuvre, as in Case Study 4 (see published video in [3]), and as in the experimental syncopes described by Lempert, Bauer and Schmidt (23). It is likely that many of the reported seizures in Rett syndrome are of this nature (59).

A boy of school age presented with an apparent recurrence of seizures. He had an early history of infantile spasms; that is, epileptic spasms with hypsarrhythmic EEG in the first year of life. The spasms remitted but he was left asymbolicthat is without the understanding of meaningand without imaginative play or social interaction. His main enjoyments seemed to be twirling or spinning dinner plates and intermittently hyperventilating and holding his breath. He was referred back because of numerous daily tonic seizures. Videorecording demonstrated a consistent stereotyped sequence. While twirling a plate, he hyperventilated, then took a deep breath in inspiration, then made a powerful Valsalva maneuver for 10 to 11 seconds, and finally with a groan he collapsed with brief tonic extension and elevation of his upper limbs. He recovered instantly. His mother said "that's a seizure, isn't it?" It was, but an anoxic seizure, not an epileptic seizure (3).

Much has been written about gastroesophageal reflux in infants, but cinematographic or videorecording or full polygraphic registration of a reflux-associated episode that might be described as a seizure has not been reported, though a true reflux episode associated with an epileptic seizure has been described (60). Nonetheless, there is a persuasively recognizable condition, the "awake apnea syndrome" (61). Having been fed within the previous hour, often following an imposed change of posture, the infant gasps, is apneic, stiffens, changes color, and may then look startled. A personal case is described in Stephenson (1).

Sandifer syndrome secondary to reflux is mentioned under Miscellaneous Neurological Events below.

An important, unusual, but difficult diagnosis relates to suffocation of a baby (usually) by the mother (62). This can be termed the active form of Munchausen syndrome by proxy, or Meadow syndrome, of the factitious epilepsy type (63) and is a classic example of fabricated or induced illness (64). In this situation, the parent repeatedly suffocates the baby by either pressing a hand or some other material over the baby's mouth, or else the mother presses the baby's face against her bosom (1) with a resultant syncope and anoxic seizure. The evolution here is much longer than in the usual cyanotic breath-holding spells (prolonged expiratory apnea), with a latency of something of the order of 2 minutes (65). Diagnosis may be exceedingly difficult, and depends on such factors as recognizing that the episodes only begin in the presence of the mother, although various other people, such as relatives or nursing or medical staff, observe the conclusion of the episodes (66). Definitive diagnosis may require covert videorecording (65,67). Transmission of the diagnosis to the family presents great difficulties. It has been found helpful to involve another experienced pediatrician, a psychiatrist, and child protection services, before discussion with the family regarding the mechanism of induction of these truly life-threatening anoxic seizures (1).

Hyperekplexia is a rare disorder (or group of disorders) that may include dramatic neonatal onset (68) with nonepileptic convulsive syncopes that may prove fatal. Insofar as effective treatment is possible by repeatedly flexing the baby (69), diagnostic awareness should be high. An early major paper on this topic (70) described a dominantly inherited disorder in which there were stiff hypertonic neonates with later pathologic startles. Some confusion has been engendered by the title of this first paper, which referred to hyperexplexia, whereas the proper Greek term is hyperekplexia (71). The consistent specific diagnostic sign of hyperekplexia is elicited by tapping the infant's nose (72). In a normal infant, nose-tapping produces a minimal response, whereas in affected children there is an obvious and reproducible startle response including head retraction. This startle may be induced over and over again. The diagnosis in sporadic cases in which the baby is stiff and tends to startle is not too difficult. More difficult is the situation in which the baby is not stiff but does have neonatal onset convulsions with severe syncope. These dramatic nonepileptic seizures may be induced by bathing but the nose-tap test is clearly positive. Also of diagnostic value is the EEG recording during a seizure. A series of what superficially may appear to be spikes appears on the EEG, but these are actually rapidly recurring muscle potentials from scalp muscle (synchronous potentials also are seen on the ECG channel) whose fire rate decreases pari passu with slowing of both EEG and ECG in the resultant severe syncope. The genetic basis for both the usual dominantly inherited variety of hyperekplexia and apparently sporadic cases is a defect in either the alpha(1) (73) or beta (74) subunits of the strychnine-sensitive glycine receptor. Whatever the variety, clonazepam remains the prophylaxis of choice (see also Chapter 13).

Although the curiously named familial rectal pain syndrome is without doubt very rare, we have had clinical contact with three families and made or seen ictal video-recordings of three children and one adult, supporting the suggestion that this unpleasant disorder is also underdiagnosed (75). Familial rectal pain syndrome is dominantly inherited, but apparently sporadic cases occur. The presenting feature is dramatic neonatal seizures. Schubert and Cracco (76) thought these might be epileptic seizures, in part because there was a favorable response to carbamazepine, but in our patients there was no independent suggestion of epilepsy, and no paroxysmal EEG discharges during many observed seizures. There were two important clues to the diagnosis. First, there were frequent striking harlequin color changes. In particular, one side of the face would turn red while the other side would turn white. Secondly, the precipitating factor for the seizures was some sort of perineal stimulation, such as wiping or cleaning. Actually the seizures appeared to be very severe syncopes, quite similar to those seen in neonatal hyperekplexia, with bradycardia and sometimes asystole, a slowing and then flattening of the EEG, and generally a life-threatening appearance. Eventually, these syncopes abated, but the adults with a similar neonatal history described continuing attacks of excruciating pain maximum in the nether regions, and precipitated by stimuli such as passing a constipated stool.

All varieties of syncope and presyncope have certainly not yet been described. For example, a common variety of apparently life-threatening event consists of initial hypoxemia of entirely unexplained mechanism (77).

Some of the disorders listed in this section may not be fundamentally different from some of the other disorders here described, particularly in the sections on vasovagal syncope and hyperventilation syncope above, but what are often called psychological mechanisms seem of more obvious importance here. A recent study (7) that looked at paroxysmal nonepileptic events (PEN) recorded on video-EEG in children and adolescents referred to a pediatric epilepsy monitoring unit found that psychological events predominated in each age group, overwhelmingly so in adolescence.

Episodes referred to as daydreams may be mistaken for epileptic or anoxic (syncopal) absences. There may be no fundamental difference from that described in the next subsection as gratification, but the subsequent conditions are more likely to lead to diagnostic difficulties.

More or less pleasurable behavior, apparently similar to masturbation, may be seen from infancy onwards, more so in preschool girls, but also in boys (78). Rhythmic hip flexion and adduction may be accompanied by a distant expression and perhaps somnolence thereafter. Manual stimulation of the genitalia does not seem necessary. The diagnosis of infantile masturbation is more difficult when the infant or young child seems unhappy during the rhythmic movements. The relative frequency of events and occurrence in specific circumstances, such as when bored or in a car seat, lends this behavior to home videotape recording. Parents prefer the term gratification (or even benign idiopathic infantile dyskinesia) to infantile masturbation, understandably. Readers are referred to Nechay et al (78) for an extensive clinical review.

Sometimes more difficult to diagnose may be the phenomenon in slightly older children, of the "television in the sky." Affected children may appear to stare into space or have unvocalized speech with imaginary individuals and perhaps seem to twitch or move one or more limbs for several minutes at a time. When there are repeated jerks or spasms, there may be confusion with epileptic infantile spasms.

In several situations children may describe experiences in which they appear to lose immediate contact with their bodies and perhaps see themselves from above. Such hallucinations have been described in epileptic seizures, anoxic seizures (20), migraine and as a "normal" phenomenon. Some of these perceptual disorders may be described as the "Alice in Wonderland phenomenon." Dissociated states have been well described by Mahowald and Schenck (79).

Panic attacks are well recognized in adults and criteria for their presence in children have been described and well reviewed (80). However, it is important to recognize that panic attacks may actually be manifestations of epileptic seizures (8183). As the latter authors emphasis, long-term video-EEG monitoring may be necessary to establish the correct diagnosis and prevent inappropriate psychiatric interventions.

Whether the term hysteria should be used is debated but self-induced nonepileptic, nonsyncopal seizures are not rare (7). Such episodes are called by various names, such as pseudoseizures, pseudoepileptic seizures, psychogenic nonepileptic seizures, nonepileptic attack disorder, or emotional attacks; none of these terms is satisfactory for every case. The sort of episodes described may crudely mimic epileptic seizures and have some resemblance to certain frontal lobe epileptic seizures but often have prominent sexual and aggressive components. They are usually recognized readily by observation and particularly by videotape observation and do not include alteration in background EEG. Some are predominantly swoons: a more or less graceful collapse without injury often into a recovery position, in some rhythmic jerking of the head, one or more limbs or trunk or pelvis predominates. In some cases incest, child sexual abuse, or other cause of posttraumatic stress disorder (PTSD) may be the etiology (84,85).

What has been called a "psychosomatic" syncope has been described in adults who collapse on head-up tilt with normal vital signs (86). This sort of response can be seen quite frequently in head-up tilt testing in children. One such child had been expelled from school because of frequent "fainting." Collapse occurred on head-up tilt without change in heart rate, blood pressure (continuously recorded by Finapres), or EEG. Simple psychotherapy was followed by prompt recovery. The differential diagnosis here includes hyperventilation syncope.

The child or teenager's insight may not be good or may fluctuate, and emotional attacks are well recognized in teenagers who also have epilepsy. The psychiatric literature has changed terms over the years (as can be seen by comparing successive versions of DSM), from hysteria to conversion hysteria to conversion disorders to dissociative states, without adding precision or clarity. A sociomedical model is useful for professionals: considering the illness "real," recognizing that it may be an inevitable response to a particular "predicament" (87,88), and allowing the patient to recover while saving face (89). We strongly recommend to readers a clear and modern view of hysteria (90).

Reassurance and encouragement, with or without simple behavioral techniques, will often work. However, in difficult cases psychiatric management should be available; even then, identification of underlying trauma or abuse is rare.

Because autistic children may not pay attention, they may find themselves at risk of EEG studies and a false diagnosis of epilepsy, especially if other paroxysmal phenomena such as parasomnias and syncopes coexist.

In some families, seizures are not induced but are invented (63). This can be termed the passive form of Meadow's syndrome.

A preschool child was supposed to be having daily seizures. These were no longer observed after admission to the ward of a children's hospital. However, when the mother was then interviewed by an adult psychiatrist in another hospital, she affirmed that the seizures were continuing in the children's hospital with the same frequency as previously.

While it is certain that all the funny turns that may occur in the daytime have not yet been properly described in the literature, it is even more likely that the disorders of the sleep process are by no means fully described. There are great intrinsic difficulties in readily determining what happens during sleep. Even ordinary visual observation may be difficult, whereas videorecording and even more so, polygraphic recording, may only be possible in exceptional cases, where episodes are very frequent. It is important to recognize that all parasomnias have not yet been described and to question carefully the origin of any episode that occurs only during sleep, even though some disorders previously thought to be parasomnias have now been found to be epileptic.

The parasomnias and neurological disorders of sleep, such as narcolepsy, may be confused with epilepsy due to their paroxysmal nature. The difficulty in differentiating epileptic and nonepileptic events is compounded by the fact that paroxysmal nonepileptic sleep events are more common in children with epilepsy or learning disabilities than in the general childhood population (6). Sleep disorders remain a largely neglected and poorly understood area in pediatrics. However, with careful attention to the timing and semeiology of events and the use of video-EEG and nocturnal polysomnography, these conditions can be classified and distinguished from epileptic seizures (see also Chapter 16).

A detailed history will distinguish most parasomnias from epileptic seizures. Parasomnias typically occur only once or twice a night. If events are occurring at a frequency of three or more times a night, the strong likelihood is that they are epileptic in nature, most likely arising from mesial or orbital frontal lobe structures. Epileptic seizures tend to occur more frequently in stage 2 sleep but may occur throughout sleep. In differentiating these events, video polysomnography is the most useful investigative tool.

Brief nocturnal arousals are normal in children. They occur typically in stage 4 non-REM sleep, 1 to 2 hours after sleep onset. They vary from normal events such as mumbling, chewing, sitting up, and staring to arousals that can be thought of as abnormal because of the disruption they cause the family. These include calm and agitated sleepwalking, and a spectrum from confusional arousals to night terrors or pavor nocturnus. The child may exhibit automatic behavior, but the events are not truly stereotyped. The affected children may be very agitated and look frightened, as if they do not recognize their parents. They are in an intermediate stage between waking and sleep, so they may respond, but not normally. They look awake and may be partially responsive but in fact are still in deep slow-wave sleep (stage 4). These events typically only occur once a night, especially 1 or 2 hours after falling asleep and nearly always in the first half of sleep. Children have no memory for them. Often they are very prolonged. Typically the events last 10 to 15 minutes before the child either wakes, or settles back to restful sleep.

By contrast, nocturnal frontal lobe epileptic seizures typically last less than 2 minutes and often occur in clusters. The distinction between NREM arousal disorders and benign partial epilepsy with affective symptoms (BPEAS), (91) and a variety of idiopathic focal epilepsies like benign Rolandic epilepsy, can be more difficult. Children arouse and look similarly wild and combative. However, the epileptic seizures are, brief, may occur while awake, in sleep do not arise particularly from stage 4 sleep, and are more likely to occur towards the end of sleep, in the early morning.

NREM arousal disorders likely represent a disordered balance between the drive to wake and the drive to sleep. They are more common in toddlers who sleep very deeply, in children who are overtired because of insufficient sleep, and in those who are unwell or on certain medications. An increased drive to wake occurs if the child has an irregular sleep schedule, is unwell, or needs environmental associations to fall asleep normally. These disorders are therefore primarily managed by reassurance, explanation, and behavioral means to establish stable sleep routines and ensure good sleep hygiene. Home videotape recording is invaluable, particularly if the camera can be left running to capture the onset of the event. It is generally true that home videotape of nocturnal events is more likely to be successful if they are nocturnal frontal lobe seizures rather than partial arousals due to the relative frequency and clustering of epileptic events.

Nightmares and sleep paralysis are the principal REM sleep disorders that may be confused with epilepsy. They are both common. Ten to twenty percent of individuals have some experience of sleep paralysis. This is a frightening experience of paralysis when waking from REM sleep without abolishing the physiological REM atonia that prevents us from "acting out" our dreams. Nightmares are usually easier to distinguish from epileptic seizures than night terrors, as the child will have a memory of both waking and of the dream, and will then move into normal wakefulness rapidly. Nocturnal epileptic seizures rarely arise out of REM sleep. Behavioral management and treatment of any comorbid medical conditions are the appropriate treatment strategies. The onset of a REM behavior disorder may rarely be the first clinical sign of a brainstem lesion, and neuroimaging may be appropriate.

SleepWake Transition Disorders

Rhythmic movement disorders such as nocturnal head banging (jactatio capitis nocturna), body rocking, and head rolling typically occur in infants and toddlers as they are trying to fall asleep. They can be present in deep sleep and in wakefulness. They are more common in children with learning disabilities. They typically remit by 5 years of age, but may persist into adult life. Management relies on good sleep hygiene and padding the headboard so the rest of the house is not wakened. Rhythmic movement disorders that are not clearly associated with the sleepwake transition state respond less well to behavioral management techniques and (rarely) medications such as benzodiazepines may be helpful.

The major importance of benign neonatal sleep myoclonus (92) is that it may be misdiagnosed as epilepsy and even treated with such heavy doses of antiepileptic medication that the neonate ends up in the intensive care unit on a ventilator (1). Recognition is easy for someone who has seen a videotape of the condition in which the baby has repetitive, usually rhythmic but possibly arrhythmic, jerks of one or more limbs only during sleep. In some instances, there is a report of the occasional jerk in the waking state but sometimes in very young infants it is difficult to tell whether the sleep state is actually present. A simple maneuvre to provoke BNSM has been described (93). These authors found that slow (1/second) rocking of the infant's crib in a head to toe direction would reproduce the myoclonus, whichin contrast to the situation in jitterinessdid not stop if the limbs were restrained. If there are still difficulties in the diagnosis, an EEG may have to be obtained during a period of jerking, either by prolonged recording or using the maneuvre just described. If using EEG, it is imperative to ensure that artifacts that result from the perhaps quite violent jerking are not misinterpreted as epileptic spike discharges, and collodion electrodes must be used (93). Although it is important for pediatricians or pediatric neurologists to have seen a videotape of as many of the various paroxysmal phenomena described in this chapter as possible, it is absolutely essential for clinical staff to have seen the appearance of benign neonatal sleep myoclonus so that the misdiagnosis of epilepsy may be avoided.

We suggest that every EEG department keep a videotape of benign neonatal sleep myoclonus so that it can be shown to worried parents whose well baby has inadvertently been referred for EEG. Technicians in our department have become skilled at recognizing these patients and will ask the neurologist to show the parents the video. This can be extremely reassuring and frequently produces what we term the "That's it!" phenomenon.

We mention BNSM in this chapter because 2-year-old children may present for epilepsy management, particularly if other paroxysmal events coexist, as in case study #6.

A 2-year-old girl was reported to have had twitching of the right hand for several weeks after the first day of life, sufficiently repetitive and vigorous to elicit treatment with phenytoin and phenobarbitone. Recent "generalized seizures" prompted referral by her pediatrician. When the mother was shown a videorecording of three other children displaying the repetitive generalized jerks of benign neonatal sleep myoclonus, she became excited and said "that's it!".

She had mentioned the right hand because it was often more obvious: the jerks had been violent, all over (four limbs) and lasting seconds. They had only been seen in sleep. The recent episodes were nonepileptic reflex syncopes (expiratory apnea or reflex asystole), but the double history had provoked the false diagnosis of secondary generalized epilepsy.

Vigevano's group (94) reported with videorecordings repetitive sleep starts in children who also had epilepsy and tetraplegic cerebral palsy. These jerks occurred repetitively at the onset of sleep, in clusters lasting several minutes, with arousal appearance on EEG but no jerk-related spike discharges. The authors emphasized the need to differentiate these sleep starts from epileptic seizures, particularly as the children also had epilepsy, so as to avoid excessive inappropriate antiepileptic medication.

We include this syndrome, though generally thought of as a condition of middle age, because it may present in childhood as an attention deficit disorder (95) and hence the possibility of misdiagnosis as absence epilepsy. Recognition is worthwhile as it tends to be exquisitely dopa sensitive (96). It is seen in children with leukemia, often as a consequence of chemotherapeutic agents, and in this situation may be responsive to benzodiazepines. In children with renal failure, it is important that iron deficiency is treated.

Narcolepsy is a disorder characterized by excessive daytime sleepiness, cataplexy (a loss of tone in response to strong emotion, typically laughter), sleep paralysis, hypnagogic hallucinations, and disturbed nighttime sleep. A third of adults describe onset before 16 years of age, about 16% before 10 years, and around 4% less than 5 years (97). Deficiency of the neurotransmitter hypocretin (also known as orexin), produced in the hypothalamus, has recently been confirmed in human narcolepsy (98). Hypocretins help mediate arousal and project to brainstem structures involved in muscle tone. It is likely that narcolepsy is primarily an autoimmune neurological disorder due to damage to the hypocretin producing system in genetically susceptible individuals. Consciousness is maintained during cataplexy even though the eyes may be closed. Diagnostic confusion may arise if several attacks of cataplexy occur one after the other and then the individual falls asleep on the floor (99). Typically, the loss of tone spreads from the face down the body. The individual maintains a degree of control, so as they collapse this often appears to occur in a series of stages rather than a sudden fall. In a personal series of six children diagnosed with narcolepsy between 1997 and 2000, four had been given a diagnosis of epilepsy: either absences because of the excessive sleepiness, myoclonic drop attacks because of the cataplexy, or partial seizures because the cataplexy was asymmetric. One child had been treated with multiple antiepileptic medications. Diagnosis rests on the recognition of the five features of the syndrome, video-recording of cataplexy if possible and practicable, and the multiple sleep latency test, provided the child is 8 years or older (100).

A complex relationship exists between epilepsy and movement disorders, the boundaries of which are difficult to define (101). They share many symptoms and are frequently confused with each other. Paroxysmal movement disorders are characterized by a variable duration of motor symptoms, usually with few if any interictal abnormalities on examination. Some children with "intermediate" exertion-related dystonia have subtle dystonia or signs of developmental dyspraxia, even on good days. The major distinguishing features between these events and epileptic seizures are the frequent presence of precipitating factors and the retention of consciousness in the paroxysmal dyskinesias and ataxias. These features may be more difficult to determine in childhood.

Historically, this group of disorders has been classified separately from the idiopathic epilepsies. Recent reports have emphasized the co-occurrence of movement disorders and epilepsy in the same family, thus suggesting that they both may have the same underlying mechanism (101104). The recognition that dysfunction of ion channels leads to cellular hyperexcitability, and that mutations in these channel proteins may be associated with both epilepsy and movement disorders, may explain this relationship (105). A credible hypothesis is that mutant ion channels are expressed in variable degrees in different central nervous system (CNS) structures and that this expression may also vary with brain development. Thus, the phenotype of a genetic ion channelopathy may include partial epileptic seizures in infancy, indicating a cortical pathology, and an episodic ataxia in childhood and adolescence, suggesting cerebellar dysfunction. It is relevant to note that many antiepileptic medications are effective treatments for paroxysmal movement disorders.

The neurological CNS channelopathies comprise a large group of disorders, including epilepsy, migraine, movement disorders and hyperekplexia, which are characterized by their paroxysmal nature (106). Why the channelopathies are paroxysmal is not well understood but is likely to relate to the vast number of complex interacting influences on ion channel function (107). These include cell membrane voltage, pH, temperature, intracellular and extracellular ligands, phosphorylation of channel proteins, electrolyte status, and the relative expression of mutant and wild-type subunits of the channel, which will vary over time.

Various complex classifications have been proposed for this group of disorders [see Fahn (108) for a summary and a history of terminology]. The most clinically relevant and simplest is used here. Most of the literature describes familial cases, which are easier to diagnose (especially once one or more affected family members are known) and are possibly more interesting to report than sporadic cases. However, it is our clinical impression that most people with paroxysmal dyskinesias are sporadic, and many cases do not fit exactly into the classical descriptions outlined.

Typically, onset is in early childhood or adolescence with episodes of choreoathetosis, dystonia, or a mixed pattern. Attacks tend to become less frequent or remit totally in adult life. Attacks last seconds to 5 minutes and are precipitated by sudden movements, change in position, or change in movement velocity (109). Getting up from a chair or getting out of a car are frequent triggers. Consciousness is retained, and some individuals may have a brief nonspecific warning or aura prior to an attack. Interictal examination is normal. Diagnosis is based on history, and a videorecording of events is invaluable. Carbamazepine is often highly effective in small doses. A family history of similar events exists in about a quarter of patients, with autosomal dominant inheritance in many families. Linkage to several overlapping but distinct loci around the pericentromeric region of chromosome 16 has been reported but the genes involved have not been identified to date (110,111). In some families, the paroxysmal dyskinesia is associated with benign familial infantile convulsions (102,104). This has been reported as the infantile convulsions and choreoathetosis syndrome (ICCA). However, the movement disorder may include paroxysmal dystonia and is, therefore, better classified as a paroxysmal dyskinesia.

Attacks are often longer in PNKD and may last 2 minutes to several hours or even 2 days. This type is sometimes referred to as paroxysmal dystonic choreoathetosis (PDC). The attacks are often markedly dystonic and occur spontaneously, although in adults alcohol, caffeine, and stress are frequent precipitants. Differentiation from epileptic seizures is easier and treatment with antiepileptic medications is less effective than in true epilepsies. Inheritance is usually autosomal dominant, and linkage has been reported to chromosome 2 (112,113).

Events occur after several minutes of exercise, usually 10 to 15 minutes or more, not at the initiation of movement, as in PKD (114). Typically, the part of the body that has been doing most exercise becomes dystonic. The abnormal movement resolves gradually with cessation of the exercise, over 5 to 30 minutes (intermediate between PKC and classical PDC). Antiepileptic medications are not generally helpful although acetazolamide has been effective in some families (114).

In BPT, infants have attacks of retro-, latero-, or torticollis that may last minutes to hours (115). In rare instances, they may last days. Typically, attacks begin in early infancy and remit by age 5. They may be triggered by movement, often in the early morning, and are heralded by irritability, pallor, vomiting, and in older children, clear ataxia. BPT is both a movement disorder and a migraine equivalent (116). Two patients with BPT in a recent series came from a family with familial hemiplegic migraine linked to a mutation in the voltage-gated calcium channel gene CACNA1A on chromosome 19 (117).

Benign paroxysmal tonic upgaze of childhood (118) typically presents in infants of less than 3 months with prolonged periods (hours to days) of sustained or intermittent upgaze deviation. Later, ataxia is associated. The episodes remit with in a few years but are associated with psychomotor retardation or language delay in up to 80% of cases (119).

Episodic ataxia type 1 (EA1) is a rare disorder caused by mutations in the voltage-gated potassium channel Kv1.1. Affected individuals have brief episodes of cerebellar ataxia lasting seconds or minutes (120). Interictal myokymia, detected clinically or by demonstration of continuous motor unit activity on EMG, is the principal diagnostic feature. As well as this paroxysmal ataxia being confused with a partial epileptic seizure, there exists a real over-representation of epilepsy in families with EA1 (103,121). The potassium channel is expressed throughout the central and peripheral nervous system. Whether the phenotype comprises ataxia, myokymia (or neuromyotonia), or epilepsy or a combination of these seems to relate to the functional consequences of the mutation and its tissue-specific developmental expression (121).

Episodic ataxia type 2 (EA2) is less frequently mistaken for epilepsy because the attacks are longer (minutes to hours), and there may be interictal cerebellar signs including eye movement control impairments. This disorder is associated with mutations in the voltage-gated calcium channel gene CACNA1A located on chromosome 19 (122). It is allelic with familial hemiplegic migraine and spinocerebellar ataxia type 6. In their pure forms, these are distinct disorders but overlap syndromes do occur. Partial seizures have been documented in familial hemiplegic migraine families, and there is a case report of a child with a de novo truncating mutation in CACNA1A who has EA2 and absence epilepsy (123,124).

Some authors regard migraine with aura as an important differential in the diagnosis of epilepsy (125). A number of conditions exist, ranging from undoubtedly varieties of definite migraine through migraine equivalents, probably having a migrainous origin to conditions in which the migraine link is more tenuous. On the whole, the more classical the migraine picture, the easier the diagnosis.

Insofar as virtually all attacks of familial hemiplegic migraine (FHM) are associated with headache and a family history of hemiplegic migraine (126), the differential diagnosis normally should not be difficult.

This is the most common of the migraine equivalents (116). Although affected preschool children are often referred to as having epilepsy, the characteristic history of anxious arrest of movement without loss of awareness and subjective vertigo or "drunking" makes the diagnosis easy. A related migraine equivalent, benign paroxysmal torticollis of infancy, has been discussed under Paroxysmal Disorders of Movement.

A 2-year-old girl had a 6-month history of episodes in which she had been dizzy, unsteady, and pale. There was no definite precipitation. She would say "Oh mummy dizzy dizzy dizzy" or "fright" and stand and cling onto her mother. At the same time something happened to her eyes. Her mother could not mime the speed of the eye movements, but recognized them as nystagmus after the child had been spun in a rotating chair (an alternative trigger would have been to elicit optokinetic nystagmus with a tape measure or drum).

Although perhaps even more rare than the better known alternating hemiplegia described in the next section, benign nocturnal alternating hemiplegia of childhood is more probably migraine related (127). Neurodevelopmentally normal young children experience recurrent attacks of hemiplegia arising from sleep. Attacks begin at about age 2 years and the course is benign.

The paroxysmal features and neurology of alternating hemiplegia of childhood are remarkable and fascinating. In their original report, Verret and Steele (128) described eight cases from the Hospital for Sick Children, Toronto; they regarded the condition as infantile onset complicated migraine. Casaer (129) only managed to include twelve cases in a multicenter European flunarizine trial. Since then, ten patients were reported from Montreal (130), a further twenty-two patients from Aicardi's group in Paris (131), and most recently forty-four patients from Boston (132). These and other figures suggest that the condition has in the past been both underdiagnosed and underreported.

The general features are well known to all pediatric neurologists, with attacks of flaccid hemiplegia on one or both sides, beginning in the first 18 months of life. This is associated with autonomic phenomena and the gradual appearance of developmental delay unsteadiness and a degree of choreoathetosis. Actually, paroxysmal hemiplegia is not the first symptom and usually the first hemiplegic attack is not noticed until after the age of 6 months or considerably later. The initial manifestations begin before the age of 6 months, often in the neonatal period. The earliest manifestation is commonly a disorder of eye movements, in particular nystagmus and strabismus. The nystagmus is paroxysmal and frequently unilateral. The strabismus may be paroxysmal also, and associated with signs of transitory internuclear ophthalmoplegia (133). Tonic and dystonic episodes also appear early in infancy, well before the first hemiplegic attack. These consist of predominantly brief and perhaps clustered tonic attacks that may be easily be mistaken for epileptic tonic seizures. These stiffenings are commonly unilateral, with some resemblance to the asymmetric tonic neck reflex, which may also be bilateral, with a degree of opisthotonus and up-deviation of the eyes. Pallor and crying or screaming and general misery tend to accompany these attacks.

Once hemiplegic episodes begin they may affect one or both sides [or even one upper limb and a contralateral lower limb as in case 3 of Casaer (129)]. Bilateral hemiplegia is associated particularly with autonomic phenomena and drooling.

Some sort of trigger precedes attacks in most affected children. Emotional factorsexcitement, bright lights, and bathing, including hot bathsare reported. The frequency of bathing as a trigger (in the bathroom, not in the sea) is probably underreportedin one family [case 3, Casaer, (129)] this regular trigger was not recognized until the child was 15 years old.

Developmental delay, ataxia, and persistent choreoathetosis develop in the majority of children, and a few develop migraine with aura (case 3, Casaer [129]; Silver and Andermann [130]).

Many paroxysmal disorders or episodic phenomena can be described as of neurologic origin, and these may be mistaken for epileptic seizures. Some more or less well-recognized examples are briefly described. It is not possible to give a complete picturemany types of events, fits, attacks, turns, or spells have surely yet to be described.

Tics, whether simple, complex, or as part of Tourette syndrome, do not usually pose diagnostic difficulty. However, if tics are frequent, as they usually are, the alternative diagnosis of an epileptic origin may be determined by recording an EEG during the tic, preferably with simultaneous videorecording.

Nonepileptic myoclonus occurs in many situations. If there is difficulty in diagnosis, EEG will determine whether the myoclonus is epileptic or not. The EEG (preferably with surface electromyography (EMG) and videorecording simultaneously) will show obvious spike discharges during epileptic myoclonus. Mention should be made of the myoclonusdystonia syndrome, if only to remark on the unexpected finding of mutations in the gene for -sarcoglycan (134,135).

Cataplexy in the narcolepsycataplexy syndrome was discussed in Derangements of the Sleep Process. Cataplexy is very rarely associated with acquired brain stem lesions. Cataplexy may be seen in Niemann-Pick type C, Norrie disease, the Prader-Willi Syndrome (136), and as an isolated familial trait. Recognition is based on the identification of emotional triggers, especially laughter, to the sudden loss of muscle tone. A cataplexy-like disorder is seen in the Coffin-Lowry syndrome (137) and described in the next section.

Early reports suggested that epilepsy was a feature of the X-linked but female manifesting Coffin-Lowry syndrome (CLS). Later publications (137) suggested that those with CLS did not have epilepsy but probably a cataplexy-like disorder triggered by the startle effect of unexpected sounds. Since then, others have recognized that reflex stiffenings may also occur (138,139) and even true epilepsy (140).

A girl with CLS but with considerably spared language ability began to have sudden collapses resembling cataplexy when she was about 7 years old. It became clear that an unexpected sound was the usual stimulus. In adulthood, the shutting of a door might only induce "cataplexy," but louder and more startling sounds also induced stiffening, sometimes prolonged. Apparently true cataplexy while telling a joke was videotaped. She also began to have nocturnal epileptic seizures without any obvious trigger.

Nonepileptic head drops are characterized by there being no difference in the speed of the initial flexion of the neck and the subsequent extension (141). These authors found that repetitive head-nods, which they described as bobs (in which the velocity of recovery matched that of descent) and in which the episodes were repeated (bobbing), were a consistent feature of nonepileptic head-drops. Defining epileptic nods as those accompanied by epileptic scalp EEG discharges, they found that head-bobbing did not occur as an epileptic phenomenon.

Childhood head tremor (142) is unlikely to be confused with epilepsy. It is possible that the condition is heterogeneous.

Functional blinking (143) should perhaps be in the psychogenic section: it is a differential diagnosis of the epileptic syndrome of eyelid myoclonia with absences. However, it is not associated with EEG discharges. Drug treatment may abolish the absence seizures and photoparoxysmal response in patients with eyelid myoclonia with absences who may then continue to have tic-like eyelid myoclonia, with or without being consciously aware of the blinking (144).

Craniocervical Junction DisordersChiari Type 1

Disorders of the craniocervical junction, particularly congenital disorders such as type 1 Chiari malformation, may be responsible for apparent syncopes that are distinctive in not being associated with EEG or ECG change. Clues are stimuli, such as coughing, which would be expected to increase downward brain herniation. Definitive diagnosis is by sagittal brain MRI.

Pyogenic meningitis must be mentioned here because the tonic or vibratory nonepileptic seizures that accompany brain swelling (as in Hemophilus influenzae meningitis) were often misdiagnosed as epilepsy and treated with repeated injections of diazepam with disastrous results [e.g., (1), case 15.46]. Although immunization should now prevent serious hemophilus infections, the situation of brain swelling and herniation is by no means confined to this disorder and can be seen with any acute rise in intracranial pressure, as with intracranial hemorrhage or decompensated hydrocephalus.

Aside from metabolic derangements in which the diagnosis is obvious, tetany is most often seen with hyperventilationsee vasovagal syncope, hyperventilation syncope, panic and anxiety attacks. In hypoparathyroidism, it is more usual to have some form of epileptic seizure than tetany.

Intermittent contortions of the neck with marked lateral flexion are occasionally seen with severe gastroesophageal reflux, either in normal or in neurologically impaired children (145).

A recent publication (146) has highlighted the occurrence of a new form of nonepileptic seizure in the first 3 months of life. Episodes of sudden stiffening occurred in wakefulness in otherwise normal children and were almost exclusively precipitated by holding the infant upright in a vertical posture. The stiffening lasted for a few seconds and was accompanied by apnea and cyanosis, often followed by crying. Whether this is completely different from the awake apnea of gastroesophageal reflux (61) is unclear.

We conclude this section on Miscellaneous Neurologic Events with a disorder that is most important in the differential diagnosis of epileptic infantile spasms. We were tempted to have placed it in Psychological Disorders next to Gratification (Including Infantile Masturbation) and Stereotypies, but as we shall see there are hints that it may sometimes be a marker of deviant nervous system development, albeit to a very mild degree.

In the first description (147) and in the most recent publications (148, 149) this has been called "benign myoclonus of early infancy", but we agree with Charlotte Dravet and her colleagues (150) that a much better term is "benign nonepileptic infantile spasms". This is because the repetitive axial and limb muscle contractions last longer than myoclonus and are easily misdiagnosed as epileptic infantile spasms (149). The distinction is that in benign nonepileptic infantile spasms the EEG is normal not only interictally, but during the runs of spasms as well.

Shuddering attacks (151) have been regarded by some authors as a separate condition that may sometimes be an early manifestation of essential tremor (152), but we agree with Kanazawa (153) that shuddering attacks and benign myoclonus of early infancy and benign nonepileptic infantile spasms are all the same thing. A difficulty may be that most physicians have not seen such video recordings which have only rarely been published (148). A case study (case study 9) illustrates the emotional harm that may be done by a precipitate diagnosis of epileptic infantile spasms or West syndrome.

An otherwise normal male infant was referred to the authors' institution at the age of 6 months, with a 1 month history of daily episodes now increasing in frequency. Now three to four times daily he would have a series of clusters of spasmodic extensions of the upper limbs with flexion of his neck and stiffening of his lower limbs. During these runs of spasms he would seem unresponsive although his eyes were open with some upward deviation of the globes. His pediatrician witnessed an episode of serial spasms and thought the appearance typical of infantile spasms (as did renowned paediatric neurologists and epileptologists when video recordings were shown at international meetings in due course). West syndrome was discussed with the parents, who formed the impression that their baby's chance of eventual normal intellect was no more than 10% and were much distressed.

It was notable on detailed questioning that episodes initially occurred only when he was in his high-chair at feeding time, and later when the boy was in his car seat, sitting in a shopping trolley, or sitting on the floor. By the time of his referral to us these serial spasms occurred during every meal, at breakfast, lunch, tea, and dinner. They were never seen when he was lying in bed or standing up. His mother felt that she could prevent or stop the episodes by clapping or talking to him. She thought they were more likely if he was exhausted or exasperated or if his food did not come in time.

Video recording of episodes in his high-chair showed repeated bowing of his head with eyes up, upper limbs outstretched with fists clenched, and lower limbs rigid. The run of spasms ceased when he was given a biscuit and he smiled. Simultaneous EEG showed no ictal complexes, and interictal EEG was also normal.

Episodes gradually lessened and ceased spontaneously at the age of 13 months. Now a schoolboy of 10 years, he has no problems at all. There is no family history of essential tremor.

Of course epileptic infantile spasms may occur in the context of normal development, but the clue to the correct diagnosis of nonepileptic infantile spasms lay in the setting and provocation. Episodes in the high-chair and in the car seat closely resemble the history in cases of infantile masturbation (78) and suggest a similar behavioural mechanism, and not a precursor of essential tremor. Such subtleties add to the richness of the Child Neurologist's work, with no prospect of closure in the foreseeable future.

In the past 40 years, those who have paid close attention to the mechanisms of seizures have recognized that the common motor seizure that is a manifestation of severe convulsive syncope is a nonepileptic seizure, the so-called anoxic seizure (see Syncopes and Anoxic Seizures). Many comments in the literature suggest that it is "common knowledge" that severe anoxia produces epileptic convulsions, or even specifically tonic-clonic epileptic seizures, but, in fact, such remarks have until recently been based on a misinterpretation of the data. Indeed, there is no published, well-documented instance of a generalized tonic-clonic epileptic seizure, as currently defined, ever having been an immediate sequel of acute anoxia, either asphyxial or ischemic. Nonetheless, true epileptic seizures as an immediate consequence of syncope have now been properly described and recorded (1,57,154). We call this phenomenon of a syncope followed by an epileptic seizure an anoxic-epileptic seizure (AES).

Most of those reported with AES were infants or young children who also had a history of reflex syncopes without an epileptic component. Most of these syncopes were reflex anoxic seizures, with reflex asystolic syncope or reflex expiratory apnea (cyanotic breath-holding spells), or syncopes which, from the description, might have been mixed breath-holding. Compulsive Valsalva maneuvers were also responsible (56).

To date, the reported induced epileptic seizures have been predominantly clonic or absence in type. Status epilepticus (155) is either common or a stimulus to the medical attendants to write a paper!