Child Neurology
7th Edition

The general appearance of the skull can suggest the presence of macrocephaly, microcephaly, or craniosynostosis. Prominence of the venous pattern might accompany
increased intracranial pressure. Flattening of the occiput is seen in many developmentally delayed youngsters. Conversely, occipital prominence can indicate the Dandy-Walker malformation complex. Biparietal enlargement suggests the presence of subdural hematomas and should raise the suspicion of child abuse. Palpation of the skull can disclose ridging of the sutures, as occurs in craniosynostosis. Biparietal foramina are usually benign and are often transmitted as a dominant trait (3). Some are due to mutations in the MSX2 gene, whereas in other families it is part of the 11p11.2 deletion syndrome (4). Prominence of the metopic suture is seen in some youngsters with developmental malformations. Percussion of the skull can reveal areas of tenderness resulting from localized osteomyelitis, an indication of an underlying brain abscess. Macewen (cracked pot) sign accompanies the separation of sutures and reflects increased intracranial pressure.

If accurately measured, serial head circumferences continue to be one of the most valuable parameters for assessing the presence of hydrocephalus or microcephaly. After multiple measurements are made with a good cloth or steel tape to ensure that the maximum circumference has been obtained, the value should be recorded on a head growth chart (Fig. I.1). Delayed head growth, particularly with evidence of arrest or slowing of head growth, reflects impaired brain growth from a variety of causes. Scalloping of the temporal fossae frequently accompanies microcephaly and suggests underdeveloped frontal and temporal lobes. Occasionally, one encounters a youngster, usually a girl, with a head circumference at or below the third percentile whose somatic measurements are commensurate and whose intellectual development is normal.

Palpation of the anterior fontanelle provides a simple way of estimating intracranial pressure. Normally, the fontanelle is slightly depressed and the pulsations are hardly felt. Auscultation of the skull using a bell stethoscope with the child in the erect position is performed over six standard listening points: globes, the temporal fossae, and the retroauricular or mastoid regions. In all cases, conduction of a cardiac murmur should be excluded. Spontaneous intracranial bruits are common in children. These are augmented by contralateral carotid compression. Wadia and Monckton (5) heard unilateral or bilateral bruits in 60% of 4- to 5-year-old children, 10% of 10-year-old children, and 45% of 15- to 16-year-old adolescents. Intracranial bruits are heard in more than 80% of patients with angiomas. Unlike benign bruits, they are accompanied by a thrill and are much louder and harsher. Intracranial bruits are heard in a variety of other conditions characterized by increased cerebral blood flow. These include anemia, thyrotoxicosis, and meningitis. Bruits also accompany hydrocephalus and some (not necessarily vascular) intracranial tumors. The technique and the history of auscultation for intracranial bruits are reviewed by Mackenzie (6).

The child’s station (i.e., posture while standing) can usually be discerned before the start of the examination. Similarly, the walking and running gaits can be seen by playing with the youngster and asking him or her to retrieve a ball and run outside of the examining room. In the course of such an informal examination, sufficient information can be obtained so that the formal testing of muscle strength is only confirmatory.

Evaluation of the motor system in a school-aged child can be done in a formal manner. Examination of selected proximal and distal muscles of the upper and lower extremities is usually sufficient. A book published by the Medical Research Council is invaluable for this purpose (21): Muscle strength is graded from 0 to 5. The following grading system has been suggested:

Muscle tone is examined by manipulating the major joints and determining the degree of resistance. In toddlers or infants, inequalities of tone to pronation and supination of the wrist, flexion and extension of the elbow, and dorsi and plantar flexion of the ankle have been found to provide more information than assessment of muscle strength or reflexes.

A sensitive test for weakness of the upper extremities is the pronator sign, in which the hand on the hypotonic side hyperpronates to palm outward as the arms are raised over the head. Additionally, the elbow may flex (Fig. I.3). In the lower extremities, weakness of the flexors of the knee can readily be demonstrated by having the child lie prone and asking the child to maintain his or her legs in flexion at right angles at the knee (Barré sign).

Coordination can be tested by applying specific tests for cerebellar function, such as having the youngster reach for and manipulate toys. One may be reluctant to use marbles for this purpose for fear that a small child might swallow them. Ataxia with tremor of the extremities can be demonstrated in the older child on finger-to-nose and heel-to-shin testing. Accentuation of the tremor as the extremity approaches the target is characteristic of cerebellar dysfunction (intention tremor). In the finger-to-nose test, the arm must be maintained abducted at the shoulder. The examiner can discover minor abnormalities by moving the finger to a different place each time. The ability to perform rapidly alternating movements can be assessed by having the child repeatedly pat the examiner’s hand, or by having the child perform rapid pronation and supination of the hands. In the lower extremities, rapid tapping of the foot serves a similar purpose. Pyramidal and extrapyramidal lesions slow rapid succession movements but leave intact the execution of each stage of the movement so that no true dissociation occurs. The heel-to-shin test is not an easy task for many youngsters to comprehend, and performance must be interpreted with regard to the child’s age and level of intelligence.

A variety of involuntary movements can be noted in the course of the examination. They may be seen when the child walks or is engaged in various purposeful acts.

Athetosis indicates an instability of posture, with slow swings of movement most marked in the distal portions of the limbs. The movements fluctuate between two extremes of posture in the hand, one of hyperextension of the fingers with pronation and flexion of the wrist and supination of the forearm and the other of intense flexion and adduction of the fingers and wrist and pronation of the forearm.

Choreiform movements refer to more rapid and jerky movements similar in their range to the athetoid movements but so fluid and continuous that the two extremes of posture are no longer evident (22). They commonly involve the muscles of the face, tongue, and proximal portions of the limbs. In children, athetosis and choreiform movements occur far more frequently as associated, rather than as isolated, phenomena.

Dystonia is characterized by fixation or relative fixation in one of the athetotic postures. When dystonia results from perinatal asphyxia, it is nearly always accompanied by other involuntary movements. The other manifestations of basal ganglia disorder (tremors and myoclonus) are usually less apparent. Tremors are rhythmic alterations in movement, whereas myoclonus is a relatively unpredictable contraction of one or more muscle groups. It can be precipitated by a variety of stimuli, particularly sudden changes in position, or by the start of voluntary movements. In addition to these movement disorders, children with dystonic cerebral palsy also exhibit sudden increases in muscle tone, often precipitated by attempts at voluntary movement (tension). These movements must be distinguished from seizures.

Small, choreiform-like movements are common in the healthy infant. They are transient; emerging at approximately 6 weeks of age, they become maximal between 9 and 12 weeks of age and taper off between 14 and 20 weeks of age. According to Prechtl and coworkers, their absence is highly predictive of neurologic abnormalities (23).

A proper sensory examination is difficult at any age, and almost impossible in an infant or toddler. Sensory modalities can be tested in the older child using a pin or preferably a tracing wheel. In infants or toddlers, abnormalities in skin temperature or in the amount of perspiration indicate the level of sensory deficit. The ulnar surface of the examiner’s hand has been found to be the most sensitive, and by moving the hand slowly up the child’s body, one can verify changes that one marks on the skin and rechecks on repeat testing.

Object discrimination can be determined in the healthy school-aged child by the use of coin, or small, familiar items such as paperclips or rubber bands.

The younger the child, the less informative are the deep tendon reflexes. Reflex inequalities are common and less reliable than inequalities of muscle tone in terms of
ascertaining the presence of an upper motor neuron lesion. The segmental levels of the major deep tendon reflexes are presented in Table I.1.

Little doubt exists that the Babinski response is the best-known sign of disturbed pyramidal tract function. To elicit it, the plantar surface of the foot is stimulated with a sharp object, such as the tip of a key, from the heel forward along the lateral border of the sole, crossing over the distal ends of the metatarsals toward the base of the great toe. Immediate dorsiflexion of the great toe and subsequent separation (fanning) of the other toes constitutes a positive response. Stimulation of the outer side of the foot is less objectionable and can be used in children who cannot tolerate the sensation of having their soles stimulated. The response is identical. An extensor plantar response must be distinguished from voluntary withdrawal, which, unlike the true Babinski response, is seen after a moment’s delay. It also must be distinguished from athetosis of the foot (striatal toe). According to Paine and Oppe, a positive response to Babinski sign is seen normally in the majority of 1-year-old children and in many up to 2 1/2 years of age (24). In the sequential examination of infants conducted by Gingold and her group, the plantar response becomes consistently flexor between 4 and 6 months of age (25).

Many eponyms, 20 according to Wartenberg (26), have been attached to the reflexes elicitable from the sole of the foot. Some, such as Rossolimo reflex, which is elicited by tapping the plantar surface of the toe and producing a stretching of the plantar flexors, are muscle stretch reflexes. Others, such as Oppenheim reflex (a firm stroke with finger and thumb down the anterior border of the tibia) or Gordon reflex (a hard squeeze of the calf muscle), are variants of Babinski response.

In the upper extremity, Hoffmann reflex is elicited by flicking the terminal phalanx of the patient’s middle finger downward between the examiner’s finger and thumb. In hyperreflexia, the thumb flexes and adducts, and the tips of the other fingers flex. Wartenberg sign is elicited by having the patient supinate the hand, slightly flexing the fingers. The examiner pronates his or her own hand and links his or her own flexed finger with the patient’s. Both flex their fingers against each other’s resistance. In pyramidal tract disease, the thumb adducts and flexes strongly, a reemergence of the forced grasp reflex.

Clonus is a regular repetitive movement of a joint elicited by a sudden stretching of the muscle and maintaining the stretch. It is most easily demonstrable at the ankle by dorsiflexion of the foot. Clonus represents increased reflex excitability. Several beats of ankle clonus can be demonstrated in some healthy newborns and in some tense older children. A sustained ankle clonus is abnormal at any age and suggests a lesion of the pyramidal tract.

Chvostek sign, a contraction of the facial muscles after percussion of the pes anserinus of the facial nerve (just anterior to the external auditory meatus), is evidence of increased irritability of the motor fibers to mechanical stimulation such as occurs in hypocalcemia (27).

Children with developmental disabilities, such as minimal brain dysfunction and attention-deficit disorders, are often found to have soft signs on neurologic examination (see Chapter 18). These represent persistence of findings considered normal at a younger age. Of the various tests designed to elicit soft signs, tandem walking, hopping on one foot, and the ability of the child to suppress overflow movements when asked to repetitively touch the index finger to the thumb have been found to be the most useful. Forward tandem gait is performed successfully by 90% of 5-year-old children; 90% of 7-year-old children also can hop in one place, and synkinesis becomes progressively suppressed between 7 and 9 years of age (28).

Evaluation for the presence of cognitive limitations is an important part of the neurologic examination of developmentally delayed youngsters and of children with ostensibly normal intelligence who are referred because of school failure. Such an examination is extremely time consuming and might require a return visit. An outline of an evaluation of intelligence, speech, and disorders of cognitive function is presented in Chapter 18. Also provided are suggestions on how to interpret psychological data.

Evaluation of posture and muscle tone is a fundamental part of the neurologic examination of infants. It involves examination of the resting posture, passive tone, and active tone. Posture is appreciated by inspecting the undressed
infant as the infant lies undisturbed. During the first few months of life, normal hypertonia of the flexors of the elbows, hips, and knees occurs. The hypertonia decreases markedly during the third month of life, first in the upper extremities and later in the lower extremities. At the same time, tone in neck and trunk increases. Between 8 and 12 months of age, a further decrease occurs in the flexor tone of the extremities together with increased extensor tone (29).

Evaluation of passive tone is accomplished by determining the resistance to passive movements of the various joints with the infant awake and not crying. Because limb tone is influenced by tonic neck reflexes, it is important to keep the child’s head straight during this part of the examination. Passive flapping of the hands and the feet provides a simple means of ascertaining muscle tone. In the upper extremity, the scarf sign is a valuable maneuver. With the infant sustained in a semireclining position, the examiner takes the infant’s hand and pulls the arm across the infant’s chest toward the opposite shoulder (Fig. I.4). The position of the elbow in relationship to the midline is noted. Hypotonia is present if the elbow passes the midline. In the lower extremity, the fall-away response serves a similar purpose. The infant is suspended by the feet, upside down, and each lower extremity is released in turn. The rapidity with which the lower extremity drops when released is noted. Normally, the extremity maintains its position for a few moments, then drops. In hypotonia, the drop occurs immediately; in hypertonia, the released lower extremity remains up.

The traction response is an excellent means of ascertaining active tone. The examiner, who should be sitting down and facing the child, places his or her thumbs in the infant’s palms and fingers around the wrists and gently pulls the infant from the supine position. In the healthy infant younger than 3 months of age, the palmar grasp reflex becomes operative, the elbows tend to flex, and the flexor muscles of the neck are stimulated to raise the head so that even in the full-term neonate the extensor and flexor tone are balanced and the head is maintained briefly in the axis of the trunk. The test is abnormal if the head is pulled passively and drops forward or if the head is maintained backward. In the former case, abnormal hypotonia of the neck and trunk muscles exists; in the latter case, abnormal hypertonia of the neck extensors exists. With abnormal hypertonia, one also might note the infant’s head to be rotated laterally and extended when the infant is in the resting prone position.

The evaluation of various primitive reflexes is an integral part of the neurologic examination of the infant. Many of the reflexes exhibited by the newborn infant also are observed in a spinal animal, one in which the spinal cord has been permanently transected. With progressive maturation, some of these reflexes disappear (Tables I.2 and I.3). This disappearance should not be construed as meaning that they are actually lost, for a reflex once acquired in the course of development is retained permanently. Rather, these reflexes, which develop during intrauterine life, are gradually suppressed as the higher cortical centers become functional.

The last part of the neurologic examination involves tests similar to those performed in older children or adults, such as the funduscopic examination and the deep tendon reflexes. These have been discussed already in the Cranial Nerves section and the Reflexes section. Though a variety of deep tendon reflexes can be elicited in the infant, they are of limited value except when they are clearly asymmetric. The triceps and brachioradialis reflexes are usually difficult to elicit during the period of neonatal flexion hypertonia. The patellar reflex is accompanied by adduction of the opposite thigh, the crossed adductor reflex. This reflex disappears by 9 to 12 months of age (24). An unsustained ankle clonus is common in the healthy neonate.

The ready availability of head ultrasound has for most practitioners relegated transillumination of the infant’s skull to the scrap heap of history. Nevertheless, when performed correctly, it remains a quick and useful test for detecting the presence of hydrocephalus, subdural effusions, porencephalic cysts, and posterior fossa cysts (32).

It is not always possible to summarize the neurologic examination result of the infant as being normal or abnormal. Instead, an intermediate group of infants exists whose examination results are suspect. The examination should be recorded as such, with the ultimate decision being left to subsequent examinations. Only some 1% of these suspect infants turn out to have gross neurologic deficits (33).

In most cases, a sample of the cerebrospinal fluid (CSF) is obtained by lumbar puncture, a procedure introduced in 1891 by Quincke as a means of reducing the increased intracranial pressure in children with hydrocephalus (34). To perform a spinal tap, the child is held in the lateral recumbent position with the spine maintained maximally flexed. In the newborn or small infant, the tap is best performed with the baby in the sitting position. A small pillow placed against the abdomen keeps the spine flexed while an assistant maintains the head in a perfect anteroposterior alignment. After cleaning the back with an antibacterial solution, the tap is performed. One may omit local anesthesia because it entails twice as much struggling than a tap performed without it. A controlled clinical trial has come to a similar conclusion (35). In one medical center, eutectic mixture of local anesthetics (EMLA) cream, a topical anesthetic of mixture of 2.5% lidocaine and 2.5% prilocaine, applied as a thick layer over the puncture site approximately 1 hour before the tap has made the procedure easier for the child and doctor. For a small child or toddler, a 22-gauge needle is preferable, whereas for the newborn, a 23-gauge needle without a stylet is optimal. Several authors have suggested the use of an atraumatic needle that has a blunt tip rather than the beveled cutting edge of the Quincke needle. Based on experience derived from adult subjects, such a needle would reduce the incidence of post–lumbar puncture headaches in older children and adolescents (36). The needle is inserted into the L3–4 or, for a newborn, whose spinal cord terminates at a
lower level than the older child’s, into the L4–5 interspace, with the bevel being maintained parallel to the longitudinal dural fibers to decrease the size of the dural tear. The needle is then pushed forward as slowly as possible. Entry into the subarachnoid space can be felt by a sudden pop. A bloody tap usually results from the needle’s going too deep and penetrating the dorsal epidural venous plexus overlying the vertebral bodies. Less often it is caused by injury to the vessels along the cauda equina. The person who holds the child in a perfect position is more important to the success of a lumbar puncture than the person who inserts the needle. Bonadio and coworkers constructed a graph relating the depth to which the needle must go to obtain clear CSF with body surface area (37).

If an opening pressure is essential, the stylet is replaced with a manometer, taking care to lose as little spinal fluid as possible. Because examination of the CSF for cells is more important than a pressure reading in most cases, a few drops of CSF should be collected at this point. An accurate recording of CSF pressure requires total relaxation of the child, a difficult and at times impossible task. The cooperative child may be released and the legs extended because falsely high pressures result when the knees are pressed against the abdomen. When the needle has been placed correctly, the fluid can be seen to move up and down the manometer with respirations. In older children, normal CSF pressure is less than 180 mm H₂O; usually it is between 110 and 150 mm H₂O. Pressure is approximately 100 mm H₂O in the newborn. Ellis and coworkers proposed that CSF pressure can be estimated by counting the number of drops that are collected through the needle during a fixed period of time, ranging from 12 to 39 seconds, depending on the temperature of the child and the gauge and the length of the needle. Their paper presents the counting period for which the number of drops counted equals the CSF pressure in centimeters of water (38). This method is not useful in cases in which the CSF viscosity is significantly increased. Queckenstedt test, compression of the jugular veins to determine the presence of a spinal subarachnoid block, has lost its usefulness. After the spinal tap is completed, the stylet, if such is used, is replaced, and the needle is given half a turn before being removed to prevent extensive spinal fluid leakage (39). This manipulation is believed to reduce the likelihood for a post–lumbar puncture headache (40).

The normal CSF is crystal clear. A cloudy fluid indicates the presence of cells. Fluid that contains more than 500 red blood cells/μL appears grossly bloody. If so, the possibility of a traumatic tap should be excluded by performing cell counts on three sequential samples of fluid. In the presence of an intracranial hemorrhage, little difference exists in the counts obtained from the first and last specimens. Performing sequential cell counts is a more reliable method of demonstrating intracranial bleeding than observing the presence of a xanthochromic fluid or crenated red cells (41). In the United States xanthochromia is generally assessed by visual inspection, spectrophotometry for the detection of CSF pigments is used almost universally in the United Kingdom (41a).

Crenation of red cells in CSF occurs promptly, whereas xanthochromia can be noted whenever contamination by red cells is heavy. In the term neonate, the mean protein concentration is 84 with a stardard deviation of ±45 mg/dL (42); in the premature neonate, it is 115 mg/dL (43). The presence of blood from any source raises the total protein by 1.5 mg/dL of fluid for every 1,000 fresh red blood cells/μL (44). The number of white cells in normal CSF is higher in infants than in children. In the study conducted by Ahmed and colleagues, the mean white cell count in normal, noninfected neonates was 7/μL with a range of 0 to 130/μL (45).

In children older than 12 months of age, normal values range up to 3 cells/μL. Whereas polymorphonuclear cells can be present in the newborn, they are not found in CSF taken from healthy children older than 12 months of age (46). Blood-contaminated CSF complicates the determination of pleocytosis. Although it is frequently stated that the ratio of white blood cells to red blood cells does not differ from that present in peripheral blood (1:300), actual determinations by several groups of workers indicate that only approximately 20% of the predicted number of white blood cells are present in CSF; consequently, recalculation is necessary (47).

The normal immunoglobulin G (IgG) values for pediatric patients were compiled by Rust and coworkers (48). In tuberculous meningitis, a cobweb can form at the bottom of the tube; staining of this material can reveal the presence of acid-fast organisms. It is appropriate to consider the complications of lumbar puncture performed under these and other circumstances.

The possibility of herniation in the presence of increased intracranial pressure must be considered not only in cases of brain tumor, but also in purulent meningitis. The incidence of this complication was 4.3% in the series of Rennick et al. of 445 infants and children with bacterial meningitis (49). Obtaining a CT scan prior to a lumbar puncture in patients with meningitis to determine whether there is increased intracranial pressure is in our opinion unwarranted and only results in delayed appropriate treatment. As Oliver and coworkers (50) and Rennick and coworkers (49) also pointed out, the CT is generally normal in children with meningitis and increased intracranial pressure and does not disclose incipient cerebral herniation. This subject is also covered in Chapter 7.

The most common complication of lumbar puncture is headache. In the experience of Raskin it was encountered in 10% of subjects younger than 19 years of age (51). Its onset is between 15 minutes and 12 days after the tap, and it persists for an average of 4 days but as long as several weeks or months (51). It is most severe with the patient upright and subsides in the recumbent position. Nausea, vomiting, and, less often, vertigo can
accompany the headache. It is generally accepted, although not proven conclusively, that the headache results from persistent leakage of CSF through a hole in the dura and arachnoid, with cerebral hypotension and a consequent stretching and displacement of pain-sensitive structures (51). In adults, but not in the pediatric population, the major factor in headache induction is the diameter of the hole made in the dura by the needle (52,53). Neither the amount of fluid removed nor the position of the patient during the puncture is important in terms of the incidence of post–lumbar puncture headache (53). If a small-gauge needle is used and CSF pressures are to be measured, the physician should allow sufficient time for equilibration.

The best current treatment of post–lumbar puncture headache is strict bed rest and the use of analgesics. Epidural injections of saline or intravenous caffeine have their advocates, but in the pediatric population these measures are generally unnecessary. Backache as a result of the trauma of the lumbar puncture is relatively common, but in children is generally not sufficiently severe to represent a major problem.

Less common complications of lumbar puncture include diplopia caused by unilateral or bilateral abducens palsy or combined fourth- and sixth-nerve palsies (54). This usually clears within a few days or weeks.

Trauma to the arachnoid and to dural vessels at the base of the vertebrae is fairly common, particularly in small infants, but unless the patient is suffering from a bleeding disorder, it is asymptomatic. An epidural hematoma and, even less likely, a subdural hematoma, and a subarachnoid hemorrhage after a lumbar puncture are extremely unusual complications (55). Frequent lumbar punctures, particularly when performed for the introduction of intrathecal medications, may result in the implantation of an epidermal tumor. Other, rarer complications, together with 789 references on complications of lumbar puncture, can be found in a review by Fredericks (56).

Cytologic studies can be performed on CSF after centrifugation and can assist in determining whether tumors have spread to the meninges or to the spinal subarachnoid space, as may occur in a medulloblastoma (57). Occasionally, the CSF contains choroid plexus and ependymal cells. These are particularly prominent in hydrocephalic infants (58).

Electroencephalography (EEG) remains the central tool for the clinical investigation of seizures and other paroxysmal disorders. The increasing importance of neuroimaging studies has circumscribed its usefulness, and many physicians call on EEG in clinical situations in which the procedure has little to offer (59).

EEG does not exclude the presence of epilepsy or organic disease, and its role in diagnosing the neuropsychiatric disorders is limited. A single normal tracing is of little value in excluding the diagnosis of epilepsy, and conversely, insignificant EEG abnormalities can accompany gross structural brain disease. Neither the technique of EEG nor the interpretation of normal and abnormal tracings are covered at this point. Instead, the reader is referred to Chapter 14.

EEG is indicated in transient alterations in cerebral function or behavior, central nervous system degenerative disorders (with photic stimulation), unexplained coma, prognosis after anoxic episodes and other acute cerebral problems, and the determination of brain death. It is of little help in developmental retardation, static encephalopathies unaccompanied by seizures, minimal brain dysfunction, attention-deficit disorders, and neuropsychiatric disorders (59).

Quantitative methods for data analysis of brain electrical activity (BEAM) can be used to compare signals from the various electrodes and so construct multicolor contour maps of brain activity. BEAM has found relatively little applicability in pediatric neurology. Even though this technique facilitates the detection of asymmetries, it remains investigational in mild to moderate head injuries, learning disabilities, and the attention-deficit disorders. No clinical application exists for quantitative EEG analysis without analysis of an accompanying routine EEG, and quantitative EEG by itself has not been proven useful in either the diagnosis or the treatment of children with learning or attention deficits (60,61,62). The issues involved in this procedure are discussed more extensively in Chapter 18.

As interest in sleep disorders has increased, the polysomnogram has become more readily available, not only in university centers, but also in clinical sleep laboratories. The procedure consists in the simultaneous recording of multiple physiologic variables during sleep. These include an EEG, electromyogram, electrocardiogram, and electro-oculogram. Additionally, respiration is recorded. The procedure is invaluable in the evaluation of sleep apnea of infancy and the diagnosis of suspected nocturnal seizures, narcolepsy, periodic movement disorders, and parasomnias. It is discussed more extensively in books by Guilleminault (63) and Ferber and Kryger (64).

A concentric needle electrode inserted into muscle records the action potentials generated by muscle activity. These are amplified and displayed on a cathode ray oscilloscope. The amplified input also can be fed into a loudspeaker.

Normal resting muscle is electrically silent except for a small amount of electrical activity produced by insertion of the needle that rapidly dies away. This is termed
the insertion potential. With slight voluntary activity, single action potentials become evident. With increasing volitional activity, motor unit discharges increase in number, and the frequency of discharge increases until an interference pattern of continuous motor activity is achieved.

The number or rate of motor unit discharge as well as the amplitude or shape of the individual discharges, can be abnormal. In establishing the diagnosis of myopathy, the duration of the discharge is more important than the amplitude inasmuch as the latter varies with age and with the muscle tested. Spontaneous discharges may be elicited from resting muscle or the insertion potential may be increased. After denervation of muscle, fibrillation potentials appear. They result from the periodic rhythmic twitching of single muscle fibers resulting from their hyperexcitability consequent to denervation. Fibrillation potentials appear on the average between 10 and 20 days after nerve section, and, as a rule, the greater the distance between the site of injury and the muscle, the longer it takes for the abnormality to develop. Spontaneous fibrillations also can be seen in hyperkalemic period paralysis, botulinus intoxication, muscular dystrophy, and some myopathies. They can be recorded in the proximal musculature of some term babies younger than 1 month of age and in some preterm infants (65).

Electromyography (EMG) is an integral part of the investigation of the patient with lower motor neuron disease. It is nonspecifically abnormal in upper motor neuron disease and in extrapyramidal disorders, and, therefore, has only limited use in these entities. High-spatial-resolution EMG, a noninvasive surface EMG technique, has been proposed for the diagnosis of pediatric neuromuscular disorders. As used in various European centers, its diagnostic validity appears to be similar to that of needle EMG (66). For a more extensive coverage of this procedure, the reader is referred to texts by Kimura (67) and Aminoff (68).

The conduction rates of motor nerves can be measured by stimulating the nerve at two points and recording the latency between each stimulus and the ensuing muscle contraction. The conduction rate depends on the patient’s age and on the nerve tested. Normal values for pediatric patients were presented by Gamstorp (69).

Nerve conduction velocities can be used to distinguish demyelination from axonal degeneration of the peripheral nerve and also from muscular disorders. In peripheral nerve demyelination, nerve conduction velocities are generally reduced, whereas they are normal in axonal degenerations and in muscular diseases (see also Chapters 3 and 16). The procedure also provides information about the distribution of peripheral nerve lesions.

Sensory nerve conduction velocity can be determined in infants and children but is a more difficult procedure. It is useful in the study of the hereditary sensory and autonomic neuropathies and some of the heredodegenerative disorders.

Evoked potentials are the brain’s response to an external stimulus. Most evoked potentials, being of low amplitude (0.120 μV), cannot be seen on routine EEG, but must be extracted from background activity by computed signal averaging after repeated stimuli. The presence or absence of one or more evoked potential waves and their latencies (the time from stimulus to wave peaks) are used in clinical interpretations.

In pediatric neurology, visual-evoked responses (VERs), brainstem auditory-evoked responses (BAERs), and somatosensory-evoked potentials (SSEPs) are the most commonly used tests.

Radiographs of the skull are hardly as valuable as they were before the advent of CT scan and MRI. Currently, they are used occasionally to detect and interpret localized lytic lesions, such as are seen in histiocytosis X, and to evaluate various congenital anomalies of the cranium, such as occur in achondroplasia, cleidocranial dysostosis, wormian bones, or hyperostosis. Plain skull radiography is clearly inferior to CT in the assessment of the child who has suffered a head injury, although, on occasion, radiography is better than CT in demonstrating a horizontal skull fracture.

In children, particularly in those between 4 and 8 years of age, prominence of convolutional markings is not a reliable indication of increased intracranial pressure. Separation of sutures is rare in children older than 10 years of age and is practically nonexistent beyond the age of 20 years.

The advent of CT and MRI revolutionized the diagnostic evaluation of the neurologic patients and eliminated the need for invasive studies such as pneumoencephalography and ventriculography.

The technique of CT was introduced by Ambrose and Hounsfield (74), who developed a scanning instrument based on theoretical work of Oldendorf (75). It permits visualization of even slight differences in the density of the intracranial contents without the need for and the complications of an invasive procedure. In essence, scattering of radiation is eliminated by the use of a thin x-ray beam so that the photon absorption by tissues can be calculated accurately using a sodium iodide crystal or solid-state detector. The x-ray tube and the detector move across the head and obtain readings of photon transmission through the head. The exact matrix depends on the particular scanning instrument used. After completion of the scan, the unit is rotated 1 degree and the process is repeated. In this fashion, the tube and the detector are rotated 180 degrees around the head. The resulting readings are fed into a computer that calculates the x-ray absorption coefficients of each of the voxels. The most recent machines are capable of acquiring multiple images at once using an array of thin x-ray detectors. These scanners are designed for spiral or helical scanning. With this technique the gantry rotates continuously as the table moves. The helical data set is then reconstructed into artificial slices for viewing. This technique shortens the scan time considerably. Scanners with the ability to acquire 64 slices are now available. This is beyond the technical need of most neuroimaging. Most scans are acquired as standard serial cuts of the head 2.5 or 5 mm thick. Multislice spiral imaging is used more commonly in the spine or with special techniques. The thin
sections of 1 mm or less required for examination of the posterior fossa, the orbit, or the perisellar region are easily performed by the new generations of scanners. The total radiation dose for a complete scan of the head is approximately 6 rads, somewhat less than that produced by a routine skull series.

The CT scanning procedure takes some 5 to 30 minutes, and because even a slight degree of motion results in an artifact, the normally active child younger than 5 years of age requires sedation to prevent movement of the head. Oral chloral hydrate, rectal sodium pentothal (3 to 6 mg/kg for children younger than 4 years of age; 1.5 to 3 mg/kg for children older than 4 years), and meperidine (Demerol; 2 mg/kg up to 50 mg) are reliable and safe agents for this purpose. Chloral hydrate is administered orally as an initial dose of 60 to 75 mg/kg, with a subsequent dose of 45 to 50 mg/kg for those children in whom the initial dose was ineffective, with the maximum 2 g or 100 mg/kg, whichever is less (76,77). Newborns may not require sedation because simple bundling might hold the child still enough during the short scan times commonplace with newer machines.

Nonionic contrast media of the iodide type (e.g., iohexol) can be used to facilitate recognition of structure and vascularity. They have for the greater part replaced ionic contrast media because of their lower incidence of anaphylactoid reactions and other complications (78). Enhancement of tumors and surrounding abnormal tissue results from a breakdown of the blood–brain barrier. Rapid scanning in combination with iodinated contrast allows the acquisition of CT perfusion scans (79). By measuring the change in hounsifield units in each voxel while a bolus of contrast passes through, one can calculate blood flow transit time and blood volume. In addition to being most useful for rapid assessment of strokes, this technique has an array of other potential uses (79). Unlike MRI perfusion, it provides actual values as opposed to relative ones and is more anatomic than single photon emission computed tomography (SPECT) studies.

A CT scan is the procedure of choice for the emergency evaluation of the child who has suffered head trauma or who is suspected of having a subarachnoid hemorrhage. The CT scan is also preferable to MRI for the detection of intracranial calcifications such as are seen in intrauterine infections, the phakomatoses, and brain tumors. It also can be used to localize and define a brain abscess and to follow its course under therapy.

Although catheter angiography is still the procedure of choice for the ultimate delineation of vascular lesions, these are readily detected by CT scans, MRI, and MR angiography. The latter procedures also provide information not obtainable by catheter angiography, such as the presence and extent of an associated hematoma, subarachnoid blood, or edema. CT angiography is a noninvasive technique that uses a rapid intravenous contrast bolus and thin spiral scans of a defined area of vascular anatomy. It is most useful in evaluating complex circle of Willis aneurysms (80). CT angiography is best performed as a dedicated study with submillimeter spiral acquisitions (81) (Fig. I.8).

A CT scan is inferior to ultrasonography for the diagnosis of many of the early intracranial complications of prematurity and for the evaluation of an infant with hydrocephalus. The CT scan is also inferior to MRI in the detection of minor central nervous system malformations, the evaluation of structural epileptogenic lesions, and demyelinating processes.

Magnetic resonance imaging is a versatile and noninvasive procedure, whose clinical use was first described in 1977 (82). It provides information on brain structure without exposure to ionizing radiation. For a discussion of the fundamental physical principles of MRI and an explanation of its terminology, subjects far beyond the scope of this text, and for a more extensive presentation of its application the reader is referred to books such as those by Lee et al. (83). This section reviews the applications of MRI of the brain and compares them with those for CT scanning. Table I.4 shows the relative efficacy of the two procedures. Generally, T1-weighted images can provide a higher resolution view of structural anatomy, especially with thin-section techniques or three-dimensional acquisitions. T2-weighted images are preferred for the detection of various other abnormalities. Fluid-attenuated inversion recovery (FLAIR) is a T2-weighted technique that suppresses the background water signal that makes abnormal findings even more conspicuous.

As is the case for CT scanning, MR imaging necessitates sedation of infants and of children who cannot be expected to remain still for the duration of the procedure. In our institution this requires placement of an intravenous line, using intramuscular ketamine for children who are
difficult to control, followed by intravenous midazolam and intravenous propofol or sevoflurane.

In the evaluation of tumors, MRI is generally superior to CT scanning for the detection and characterization of posterior fossa lesions, particularly when these are isodense and thus require contrast infusion for their delineation. It is also superior for the detection and delineation of low-grade tumors and provides a more precise definition of associated features, such as mass effects, hemorrhage, and edema. In a brainstem glioma, for instance, sagittal scans demonstrate the rostrocaudal extent of tumor, and tumors involving the sella or the chiasmatic cistern are more readily delineated by MRI than CT scanning because interference from surrounding bone limits the accuracy of CT scanning. The infusion of chelated gadolinium (i.e., gadopentetate dimeglumine) provides contrast enhancement for intracranial lesions with abnormal vascularity, such as the more malignant tumors, and detects lesions that disrupt the blood–brain barrier. These are best seen on T1-weighted images (84). Enhanced MRI facilitates the distinction between tumor and edema and between viable tissue and necrotic tissue. It also delineates any leptomeningeal spread of metastases and the spinal cord metastases so common in medulloblastomas. Gadolinium-contrast infusions also have been used to delineate the cerebral lesions of neurofibromatosis and tuberous sclerosis. Gadolinium contrast has proven to be an exceedingly safe material, with a complication rate of less than 1 in 10,000. The use of neuroimaging in the diagnosis of central nervous system tumors is discussed more extensively in Chapter 11.

In spinal cord tumors, MRI is the procedure of choice. It is far superior in examination of inherent bone disease and in evaluation of the bone marrow.

In addition, MRI, by providing good gray–white differentiation, offers valuable information on the status of myelin in demyelinating diseases and during normal and abnormal development. When T2-weighted images are viewed, white matter appears lighter (higher signal intensity) than gray matter for the first 6 to 7 months of life. Between 8 and 12 months of age, the signal intensities from white and gray matter are approximately equal (Fig. I.9).

Thereafter, the brain has the adult appearance with gray matter lighter than white matter (85,86). This progression is delayed frequently in developmentally retarded youngsters (see Fig. I.9). MRI of the brain in preterm infants younger than 30 weeks’ gestation demonstrates the presence of the germinal matrix at the margins of the lateral ventricles as small areas of high signal intensity on T1-weighted images and low signal intensity on T2-weighted images (87,88). The MRI also can be used to demonstrate the maturation of the cortical sulci (see Chapter 5).

For the evaluation of developmentally delayed children, particularly when anomalies of cortical architecture are suspected, MRI is far superior to CT scanning in that it can depict areas of micropolygyria, lissencephaly, or heterotopic gray matter. Additionally, the quality of images is superior to that offered by CT scans.

MRI is the procedure of choice for the evaluation of patients with refractory complex or simple partial seizures and is preferred to CT scanning in demonstrating abnormalities at the cervicomedullary junction and the cervical spinal cord. The procedure is, therefore, used as a follow-up study for the patient who has sustained cervical trauma or who is suspected of having an Arnold-Chiari malformation.

MRI has been used to map the distribution of brain iron. In the healthy brain, the concentration of iron is maximal in the globus pallidus, caudate nucleus, and putamen. It increases in the Hallervorden-Spatz syndrome, making MRI useful in its diagnosis.

Finally, in choosing between CT scanning and MRI, the physician should not forget that the critically ill patient with a variety of infusions and requiring respiratory support cannot be managed properly in many MRI units and that the procedure costs 20% to 300% more than CT scans (88).

The frequency of incidental findings on MR imaging in a normal population is significant. In a study conducted on 1,000 healthy volunteers, 18% of brain MRI demonstrated abnormal findings, the vast majority of which did not require follow-up evaluation (89).

In images obtained with diffusion-weighted MRI, contrast depends on differences in the molecular motion of water. Acute strokes, such as occur in sickle cell disease, and hypoxic-ischemic changes in mature brain are detected earlier by this technique than by conventional MRI (90,91). The advantages of this technique in neonatal hypoxic-ischemic encephalopathy are not as clear (92). Several methods, notably FLAIR pulse sequences, have been used to reduce CSF signal and produce heavy T2 weighting and thus provide additional anatomic detail. These are discussed in detail by Oatridge and colleagues (93). In our institution, a combination of FLAIR and diffusion-weighted images is used to evaluate the patient with hyperacute infarction (within 3 hours of onset).

Whereas diffusion-weighted MRI examinations use the directionally random motion of water molecules that normally occurs, DTI takes advantage of the directional (anisotropic) motion of water molecules in white matter tracts (94). The tensor is the mathematical expression of the directional movement. DTI allows the plotting of white matter tracts throughout the brain. This technique is finding application in a wide array of situations, including abnormalities in brain development, hypoxic ischemic disease, and trauma (95,96) (Fig. I.10).

Magnetic resonance angiography (97) is a noninvasive technique that allows visualization of blood vessels. In those in the pediatric age group, who have a higher cerebral blood flow than adults, the large cerebral arteries and their major branches are routinely visualized. This procedure has replaced traditional angiography for most purposes and has proved helpful in evaluating dural sinuses and cerebral aneurysms and in studying arterial and venous components of an arteriovenous malformation. In particular, it is a noninvasive diagnostic tool for the evaluation of children with vascular accidents. Some centers use gadolinium enhancement to visualize veins and arteries with slower blood flow (98,99) (see Chapter 13).

Magnetic resonance spectroscopy (MRS) is performed using the same magnets and computers as conventional MRI. Proton MRS is used more widely than ³¹P-MRS. Unlike MRI, MRS provides information on the cerebral metabolites and some neurotransmitters in one or more small regions of interest (voxels). The major metabolites that can be detected by proton MRS include N-acetyl compounds, primarily N-acetylaspartate (NAA); creatine (including phosphocreatine and its precursor, creatine); and choline-containing compounds, including free choline, phosphoryl, and glycerophosphoryl choline. NAA is a neuronal marker, whereas the choline compounds are released in the course of membrane disruption. Proton MRS also can be used to determine the concentration of lactate, which accumulates as a result of tissue damage and consequent anaerobic metabolism (100,101,102). It is the most reliable procedure for diagnosinge defects in creatine transport (103). Neurotransmitters, such as gamma-aminobutyric acid and glutamate, also can be estimated using proton MRS. In neonates, the dominant peaks are
the choline-containing compounds and myo-inositol. With maturation, NAA increases, so that by 4 months of age it becomes the major metabolic peak. As a consequence, the NAA/choline and NAA/creatine ratios increase rapidly with maturation, whereas the choline/creatine ratio decreases (100).

Concentrations of adenosine triphosphate, phosphocreatine, and some of the other high-energy phosphates involved in cellular energetics can be assessed using ³¹P-MRS.

Spectra can be acquired within 1 hour, and changes in intracellular pH and metabolites can be followed. Proton and phosphorus MRS has been used in the evaluation of muscle diseases, localization of epileptic foci, evaluation of the extent of post-traumatic lesions, classification of brain tumors, and diagnosis of the various mitochondrial disorders, leukodystrophies, and other demyelinating disorders. These techniques also have been used to determine the extent, timing, and prognosis of perinatal asphyxia (101,104).

Whereas CT perfusion uses the x-ray absorption of iodine to map perfusion, a bolus of gadolinium is used for MRI perfusion. The most frequently used technique involves signal loss caused by the influx of contrast and rapid sequences. Perfusion MRI can be used to assess the hemodynamic component in childhood CNS disease related to neoplasms and complications from their therapy, cerebrovascular occlusive disease, childhood CNS arteriopathies, and trauma. It has advantage over CT in that the entire brain can be scanned. As with CT perfusion, blood flow timing and blood volumes can be mapped (105).

As is more extensively discussed in Chapter 6, ultrasonography is widely used for the recognition of intracranial hemorrhage in the newborn and for the detection of a variety of nonhemorrhagic lesions, notably intracranial tumors, hydrocephalus, periventricular leukomalacia, and polycystic encephalomalacia (106). Additionally, ultrasonography detects areas of calcification, such as those caused by cytomegalovirus or Toxoplasma infections. Because the procedure is performed through the open anterior fontanelle, its accuracy decreases with the decreasing size of the fontanelle.

Positron emission tomography (PET) enables one to detect localized functional abnormalities of the brain. It is based on the emission by certain unstable isotopes of positrons (positively charged electrons), which, after brief passage through tissue, collide with negatively charged electrons and emit energy that can be localized by tomography. Isotopes of carbon, nitrogen, oxygen, and fluorine have been used for PET scanning. All have short half-lives and are generally prepared by an on-site or
nearby cyclotron. ¹⁸F-labeled 2-fluorodeoxyglucose is particularly useful for measuring transport and phosphorylation of glucose in that it is not metabolized beyond deoxyglucose-6-phosphate and, therefore, remains within the brain. ¹⁸F-labeled 2-fluorodeoxyglucose allows accurate measurements of cerebral blood flow and of local oxygen and glucose metabolism. Because the hallmark of an epileptic focus is an interictal area of reduced cerebral glucose metabolism, the PET scan has become invaluable in the assessment of candidates for surgical therapy of epilepsy (107). The PET scan is being used also for the early diagnosis of Huntington disease, the differential diagnoses of dementias, and the differentiation between radiation necrosis and tumor recurrence. Specific ligands have been used to investigate dopamine, opiate, and benzodiazepine receptor binding in various movement disorders. For the most part, these studies have not had any clinical applicability in pediatric cases. Finally, the PET scan has shown considerable promise in the study of the functional development of the normal and diseased human brain (108).

The so-called poor man’s PET scan, SPECT depends on the gamma-ray emission of certain neutral lipophilic isotopes, notably ⁹⁹Tc, to measure regional cerebral blood flow. In the pediatric population, SPECT finds considerable use for the study of refractory epilepsies. In essence, the results with SPECT parallel those with PET scanning; generalized hyperperfusion is seen during a seizure in the generalized epilepsies, and regional hyperperfusion is seen for a few minutes before and during a seizure in an area that corresponds to the epileptic focus in partial seizures (109). Ictal hyperperfusion extends for approximately 10 minutes into the postictal period with hyperperfused surround that often is extensive (110). Interictally, a large proportion of patients show areas of hypoperfusion corresponding to the epileptic focus. For the purpose of localizing an epileptic focus, interictal hypoperfusion is less reliable than the data obtained by a combination of SPECT and MRI (109).

The simplicity and economy of SPECT are its biggest advantages. However, several drawbacks to this technique exist. For one, SPECT tracers are not natural biological molecules, and their physiologic behavior is not fully understood. Spatial resolution of SPECT is poorer than resolution by PET, and SPECT gives no information with respect to brain metabolism, nor does it permit quantitation of blood flow. Because in most instances brain metabolism and cerebral blood flow are closely linked and the isotopes used in SPECT do not require on-site production, this procedure has an important role in the evaluation of the difficult seizure patient (111). At present, no isotopes of carbon, oxygen, or nitrogen with usable half-lives and energies exist; this procedure has been therefore mainly limited to the study of cerebral perfusion (112). Newer pharmaceuticals, notably ^99mTc-methoxyisobutylisonitrile, have been used to detect the metabolic activity of tumors (113).

In our institution, the clinical applications of functional MRI are principally in the preoperative localization of the sensorimotor, visual, and language cortex in patients with intracranial tumors (114,115). Functional MRI allows the identification of cortical activation by examination of the oxygen level of the cortical veins. Oxygenated hemoglobin has different signal characteristics than hemoglobin. Blood flow to an actively functioning cortex provides more oxygen to neurons than is consumed by them, and as a result, the oxygen level in the corresponding veins increases.

The visualization of cerebral blood vessels by the injection of radiopaque dyes was introduced by Moniz in 1927 (116). Various techniques are used for arteriography in children, most commonly cannulation of the brachial or femoral arteries under direct visualization. The exact amount injected depends on the size of the child, but, generally, nonionic contrast medium is injected directly. With the small catheters that are now available, selective angiography of the cerebral vessels can be performed on even the smallest infant (117). Lateral and anteroposterior views of the skull are taken simultaneously or with successive injections. The principal indication for the procedure is in the evaluation of vascular abnormalities, including arteriovenous malformations, aneurysms, and occlusive vascular disease (see Chapter 13). Tumors of the cerebral hemispheres are localized by the distortion of the normal vascular patterns of the carotid arterial or venous system or by the presence of abnormal vasculature. Although CT and MRI provide first-line evidence of the size and location of a brain tumor, arteriography using digital subtraction techniques offers excellent confirmatory data in terms of blood supply to the mass lesion.

Current angiographic techniques have permitted the development of interventional neuroradiology. Percutaneous endovascular procedures have become the major mode of treatment for a wide array of vascular lesions, including malformations of the vein of Galen, arteriovenous malformations, fistulas, and intracerebral aneurysms (see Chapter 13).

Spinal radiographs and CT are important in the management of the youngster who has been subjected to craniocerebral or spinal trauma. In interpreting spine films, the pediatric neurologist must consider a number of relatively common normal variations. These include the forward subluxation of C2 on C3, a variety of anomalies of the odontoid process, including the presence of an accessory ossicle, and the occipitalization of C1 (118).

MRI is the optimal study for the delineation of spinal cord tumors, diastematomyelia, the tethered cord, and various spinal cord anomalies that are reviewed in Chapter 5.

Because these procedures are of purely historical interest, a brief review suffices. In 1918, Dandy introduced the technique of visualizing the cerebral ventricles and subarachnoid spaces by the injection of air into the lumbar spinal canal (119).

When this technique was used, localization of intracranial neoplasms relied on the presence of abnormalities in ventricular size and location. The morbidity of the procedure was considerable. When air was used as the contrast medium, its introduction into the ventricular system almost invariably resulted in headache and vomiting. Less often, hypotensive episodes and focal or generalized seizures and CSF pleocytosis occurred.