I need some information

Clinical neurophysiology
Shahram Khoshbin, MD

INTRODUCTION — The brain normally produces low voltage electrical activity. This can be measured via the electroencephalogram (EEG), which is ordinarily recorded from the scalp with small surface electrodes. The precise origin of this electrical activity is unknown, but the prevailing belief is that most of the activity represents dendritic synaptic potentials in cortical pyramidal cells [1]. Since there are normal EEG recordings characteristic of certain ages and states of consciousness, it is possible to recognize generalized malfunction of the brain with this modality, as well as localized or paroxysmal abnormalities.

The body also produces electrical activity outside the central nervous system, which can be measured in health and disease. Such measurements, performed via peripheral nerve conduction studies, electromyelograms, and neuromuscular function studies, can be used for diagnostic and prognostic purposes.

A review of clinical neurophysiology is presented here. Detailed discussions of the use of the various modalities described in this topic review are presented separately in topic reviews relating to specific diseases.

ELECTROENCEPHALOGRAM — With the EEG, electrical activity is recorded from many different standard sites on the scalp according to the international (10 to 20) electrode placement system (show figure 1). The nasopharyngeal lead is a long electrode that is passed through the nose and rests on the back of the throat near the mesial aspect of the temporal lobe [2]. Since the typical EEG machine has 16 to 20 channels, a series of different 16 to 20 groups of electrode pairs are evaluated. Recording electrical activity requires measurement of voltage between two electrode sites. It is impossible to record from all possible pairs of electrodes at the same time.

There are two different styles of recording: In the referential method (previously referred to as "monopolar"), a series of different electrodes are all referred to the same electrode (eg, the "references"), which is presumed to be relatively electrically inactive (this is similar to the limb leads of an electrocardiogram). Commonly used references are the ears (A1 and A2), vertex (C2) or a non-cephalic reference, such as a "balanced neck-chest" system. With the bipolar method (which utilizes leads that are similar to the chest leads of an electrocardiogram), a series of electrodes in a line are recorded serially as successive pairs. The first recording would be from the first and second electrodes, and the second recording would be from the second and third electrodes, and so on. Use of the different montages given various view of the electrical activity at different parts of the brain.

The electrical activity from any electrode pair can be described in terms of amplitude and frequency. Amplitude ranges from 5 µV to 200 µV. Frequency of EEG activity ranges from 0 Hz to approximately 20 Hz. The frequencies are described by Greek letters: Delta — 0 to 4 Hz Theta — 4 to 8 Hz Alpha — 8 to 12 Hz Beta — More than 12 Hz

Normal EEG findings — In the normal awake adult with eyes closed, there is a prominent alpha rhythm observed in the posterior part of the head (show figure 2). The amplitude of the alpha falls off anteriorly, which is frequently replaced by some low voltage beta activity. Often, a little low voltage theta activity is observed in fronto-central or temporal regions.

The alpha rhythm disappears (or blocks) when the eyes open. This rhythm (which is prominently posterior, blocks eye opening, and is usually in the alpha frequency range) is frequently called the alpha rhythm. However, since the alpha rhythm occasionally may not be in the alpha frequency range and some alpha frequency activity in the EEG may not be the alpha rhythm, semantic confusion sometimes arises concerning the term "alpha." Thus, "alpha" has been associated with alertness and normality of level of consciousness.

With drowsiness (which is stage I sleep), the alpha rhythm gradually disappears, fronto-central beta activity may become more prominent and fronto-central-temporal theta activity becomes most significant. As sleep becomes deeper, high voltage single or complex theta or delta waves, which are called vertex sharp waves, appear centrally. Stage II sleep is characterized by increased numbers of vertex sharp waves and centrally predominant runs of sinusoidal 12 to 14 Hz activity called sleep spindles occur. Deeper sleep, characterized by progressively more and higher voltage theta and delta activity, is not usually seen in routine EEG recordings.

In routine EEG studies, some "activations" are employed to bring out abnormalities not apparent in the record without such stimulations. These include the following: Three minutes of hyperventilation Flashing a strobe light at different frequencies Drugs, particularly short-acting barbiturates. These can be administered in certain circumstances, particularly to activate epileptic activity (eg, Brevital-EEG) (show figure 3) [3].

Abnormal EEG findings — Abnormalities of the EEG are either focal (only one area of the brain) or generalized (whole brain). Additional abnormalities are either continuous or intermittent. An abnormality that suddenly appears and disappears is called paroxysmal.

The accurate interpretation of EEGs requires significant experience. There are a wide variety of normal and variant wave forms as well as artifacts that must be recognized.

Slow wave abnormalities — Increased slow activity, meaning the presence of theta and delta activity, in a recording performed when awake is virtually always abnormal. Focal delta activity is usually irregular in configuration and is termed polymorphic delta activity (PDA) (show figure 4). PDA is usually indicative of a focal brain lesion.

Another type of delta activity is intermittent, frontally-predominant, and rhythmic; this is called FIRDA (frontal intermittent rhythmic delta activity) (show figure 5), and is indicative of increased intracranial pressure in young people. It is a less specific sign of brain abnormality in the elderly. Encephalopathy — Generalized theta and delta activity is a sign of an encephalopathy. As a general rule, EEG is a sensitive test for encephalopathies. Different patterns of slow wave are associated with different types of encephalopathies [4]. In the general practice of medicine, the evaluation of encephalopathy is the most important use of the EEG, rather than diagnosis of seizure disorders.

The most common pattern with metabolic encephalopathies is generalized slowing with moderate or high wave amplitude (show figure 6) [4]. One frequent pattern observed with hepatic and renal encephalopathies, which is sometimes mistaken for epileptogenic activity, is the so-called "liver waves" or "triphasic delta waves" (show figure 7) [5]. With hypoxic encephalopathies, different degrees of slow waves with usually very low amplitudes are observed [6].

Increased beta activity is frequently due to a sedative drug or any centrally active compound, including most depressants, neuroleptics, benzodiazepines, or even alcohol and "illicit" substances (show figure 3). In general, a toxic encephalopathy is indicated by generalized slowing and beta activity.

With infectious encephalopathies, there is an admixture of slow activity and the presence of epileptogenic activity as discussed below. In herpes encephalitis, an EEG pattern is identified, which is referred to as periodic lateralized epileptic discharges or PLEDs (show figure . (See "Herpes simplex virus type 1 encephalitis").

Epileptogenic abnormalities — EEG is particularly useful in the analysis of seizure disorders. Paroxysmal abnormalities are common between overt seizures (interictally) as well as during seizures (ictally). Such abnormalities include spike and sharp waves: A spike is a single wave that stands out from background activity and has a duration of less than 80 milliseconds. A sharp wave is similar to a spike but has a duration of more than 80 milliseconds.

A spike or sharp wave is often followed by a slow wave, and spikes and slow waves can alternate at frequencies from 2 to 5 Hz.

Epileptic paroxysmal abnormalities can be generalized or focal. The classic generalized abnormality is the 3 Hz spike and wave pattern which underlies the petit mal absence attack in children (show figure 9). A typical focal abnormality is a focal single spike followed by a slow wave (show figure 10). This abnormality can be seen in focal epilepsy or secondary generalized epilepsies.

Occasionally, the abnormal activity spreads to the other hemisphere. If the abnormal electrical activity spreads rapidly to the entire brain, the EEG will be indistinguishable from that of a generalized seizure (show figure 11) [7,8].

Activations, such as hyperventilation, photic stimulation, sleep, sleep deprivation, and drugs, are useful in inducing epileptic activity and increasing test sensitivity [9]. Since it is possible to miss infrequent epileptic activity, multiple recordings are frequently helpful.

The relationship of any of these abnormalities to the particular patient is complex. As an example, the presence of paroxysmal activity may or may not mean that the problem is related to epilepsy; the final determination may well rest upon the overall clinical picture or the results of a therapeutic trial. The clinician should resist the temptation to consider the results of an EEG independent of other clinical information. In particular, a normal EEG does not exclude epilepsy, since (to cite an extreme case) the EEG may be normal during a focal seizure that is observed clinically.

Test characteristics — The sensitivity and specificity of EEGs in seizure disorders vary widely across studies. A meta-analysis of 25 studies found that specificity ranged from 0.13 to 0.99 and sensitivity varied from 0.20 to 0.91 due mainly to large interreader variation [10].

EVOKED POTENTIALS — A stimulus in any sensory modality, whether visual, auditory, or somatosensory, produces a change in the EEG. However, compared to the background EEG, the change is usually small in magnitude, with the exact configuration critically dependent upon the nature of the stimulus and the site of recording on the scalp. The evoked potential is also dependent upon and time-locked to the stimulus, and, in order to see it, the stimulus must be repeated many times and the EEG averaged (show figure 12) [2,11,12].

Evoked potentials can be utilized to test the integrity of a pathway in the CNS. The most common current use of evoked potentials is to test the speed of conduction in a particular pathway. As an example, multiple sclerosis is a disease of central myelin; if myelin is damaged, conduction is slow and evoked potentials are delayed. Since many multiple sclerosis plaques are clinically silent but can be detected with this electrical test, evoked potentials are quite useful in making the diagnosis of multiple sclerosis [13]. (See "Diagnosis of multiple sclerosis in adults").

Visual evoked potentials — Visual evoked potentials (VEPs), which were the first to become popular, are ordinarily obtained by repetitively alternating the black and white squares of a checkerboard. Each eye is stimulated individually and responses are measured from the occipital area of the scalp. Normally, the major wave is a large positive wave at about 100 millisecond latency. The wave is delayed with multiple sclerosis or optic neuritis (show figure 13).

Such slowing is not specific for multiple sclerosis. Other conditions include ocular conditions (such as glaucoma), compressive lesions of the optic nerve (such as pituitary lesions), and pathological conditions of the optic radiations or occipital cortex.

Auditory evoked potentials — Auditory stimulation produces very complex waveforms. Stimulation with brief clicks produces six small waves in the first 10 milliseconds.

The sources of this electrical activity are in serial ascending structures in the brain stem. Thus, since it is possible to study the integrity of the brain stem with these waves, the test has been used to assess "brain stem death" in cases suspected of "brain death" [11,14]. These waves may also be delayed in multiple sclerosis (show figure 14).

Somatosensory evoked potentials — Somatosensory evoked potentials (SEPs) are the averaged electrical responses in the central nervous system to somatosensory stimulation. As with sensory nerve action potentials (SNAPs) in the peripheral nervous system (see below), most components of SEPs represent activity carried in the large sensory fibers of the dorsal column, medial lemniscus — primary sensorimotor cortex pathway. Testing the speed of conduction in this pathway permits an assessment of the integrity of the somatosensory system.

SEPs from the upper extremity are commonly produced by stimulation of the median nerve at the wrist and are best recorded from a site on the contralateral parietal area. The cerebral SEP to this type of stimulation was the first EP to be discovered. SEPs from the lower extremity are produced by stimulation of the posterior tibial nerve at the ankle or the peroneal nerve at the fibular head; they are recorded best at the vertex of the head (show figure 15).

It is possible to localize a lesion in the somatosensory pathway by using short latency SEPs from subcortical structures. Although several systems of electrode placement can be used, the one that appears to produce potentials of greatest amplitude is based upon an active electrode placed over the cervical spine and an "inactive" reference site, such as vertex of the head. There are four components that can be identified, all negative peaks, which are referred to as N9, N11, N13, and N14 [15]: The N9 component clearly originates from the brachial plexus and is best recorded with electrodes placed directly over the plexus itself. The N13 component, which originates from the dorsal column nuclei, is the largest component and is occasionally the only peak identified.

By stimulating leg nerves, it is possible to obtain EPs at all levels of the neuraxis, including those over the spinal cord.

NERVE CONDUCTION STUDIES — A number of modalities are used to evaluate the integrity of peripheral nerves. These principally include sensory and motor nerve conduction studies [16].

Sensory nerve conduction — The cell bodies of sensory neurons are located in the dorsal root ganglia. Each neuron has a central process entering the spinal cord through the dorsal horn and a peripheral process connecting to a sensory receptor in the skin or deep tissues of the limb. The receptors transduce somatosensory stimuli into electrical potentials that eventually give rise to action potentials in the axons; these are transmitted along the peripheral process to the central process. This is called a sensory nerve action potential (SNAPs). These studies can be done orthodromically in the direction of conduction or antidromically in the distal part of major peripheral nerves (show figure 16).

There are a variety of functional types of sensory neurons, each with a characteristic spectrum of axonal diameters. Neurons are myelinated or unmyelinated, but the unmyelinated fibers cannot be routinely measured. Many sensory axons with differing function and size coexist in sensory nerves and with motor axons in mixed nerves.

The two goals of sensory nerve conduction studies are the assessment of: The number of functioning axons (estimated by measuring the amplitude of SNAP). The state of myelin of these axons (estimated by the conduction velocity of SNAP).

With the usual sensory nerve conduction study, all of the axons in a sensory nerve are activated with a pulse of electric current. Action potentials travel along the nerve and the electric field is recorded at a site distant from the site of stimulation. Since each axon makes a contribution to the magnitude of the electrical field, the amplitude of the recorded sensory action potential reflects the number of functioning axons. Utilizing the distance between the site of stimulation and the site of recording and the time between stimulation and the arrival of the action potentials at the recording site, it is possible to calculate a conduction velocity, which reflects the quality of myelination of the axons.

In patients with axonal degeneration neuropathies, the primary feature is reduced sensory action potential amplitudes; this can be observed, for example, in diabetic neuropathy. (See "Clinical manifestations and diagnosis of diabetic polyneuropathy"). The conduction velocity may be slightly slowed, but only to the extent that the largest axons are gone; in this setting, the measured conduction velocity reflects the velocity of the largest remaining axons. In addition, slowing of conduction is the primary feature in demyelinating neuropathies, such as Guillain-Barré Syndrome, familial neuropathies, and in compression and entrapment neuropathies, such as carpal tunnel syndrome. (See "Clinical manifestations and diagnosis of carpal tunnel syndrome").

In radiculopathies, both the sensory action potential amplitudes and conduction velocities are normal. This is because the lesion is virtually always proximal to the dorsal root ganglion, and the cell body and its peripheral process remain normal. Sensory action potentials similarly remain normal with lesions of the central nervous system (show figure 17).

Motor nerve conduction — There are significant differences between sensory and motor nerve conduction that depend in large part upon differences of anatomy. Motor neurons have cell bodies in the anterior horn of the spinal cord and send their axons to innervate muscle fibers. Motor axons are always intertwined with sensory axons. Unlike pure sensory nerves (eg, the sural nerve), there are no pure motor nerves. Thus, the electrically stimulated compound action potential of any nerve containing motor fibers is really a mixed nerve action potential. It is therefore impossible to deduce the number of functioning motor axons via the measurement of nerve action potential amplitude.

However, it is possible to study motor nerve axons separate from sensory axons by electrically stimulating the nerve and recording from the muscle fibers it innervates. The amplitude of the compound muscle action potential is very much larger than the nerve action potential since each motor axon typically innervates hundreds of muscle fibers. The amplitude of the muscle action potential is indicative of the number of activated muscle fibers.

As previously mentioned, however, the amplitude is not indicative of the number of motor axons in the nerve. As an example, the number of axons can be diminished and the action potential can remain normal if the process of collateral reinnervation of muscle fibers by the remaining axons has been complete. Conversely, the number of axons can be normal and the action potential diminished if there is diminished synaptic transmission at the neuromuscular junction or if there is loss of muscle fibers. If a neuropathy progresses and collateral reinnervation fails to keep pace, the size of the muscle action potential declines.

Motor neuron conduction velocity — Synaptic transmission between nerve and muscle is required to produce a muscle action potential. The delay associated with this process, which is referred to as the distal motor latency, prevents direct calculation of the velocity of motor nerve conduction from a single stimulus location. More specifically, distal motor latency is the total time required for the following events to sequentially occur: The motor nerve action potential to travel down the terminal branches of the axon The release of acetylcholine into the neuromuscular junction, resulting in the end plate potential The generation of a muscle action potential Depending upon the position of the recording electrodes, the time for the muscle action potential to propagate to the recording electrodes

Clearly, a calculation of the conduction velocity for this process is not similar to that for sensory nerves. However, if the nerve is stimulated supramaximally in two places, virtually identical muscle action potentials will result; the major difference is the different latencies from the time of stimulation. In turn, the difference in the latencies is due to the difference in the distances from the sites of stimulation to the muscle. Dividing the distances between the two stimulus sites by the difference in the travel times produces a conduction velocity for the segment of nerve between the two sites of stimulation. As with the sensory action potential, measurements of the muscle action potential are ordinarily made to the time of onset; hence, the calculated conduction velocity refers to the fastest (and largest) axons in the nerve (show figure 18).

With axonal degeneration neuropathies (the most common variety of generalized neuropathies), motor nerve conduction studies are not significantly abnormal until the process is moderately advanced. Total reliance upon motor nerve conduction with the exclusion of sensory nerve conduction (as practiced in many laboratories) results in the failure to detect many significant neuropathies. There is slight slowing of conduction velocity and prolongation of the distal motor latency to the extent that the largest axons are lost. There may be loss of action potential amplitude when the process is advanced.

A focal neurapraxic lesion (eg, carpal tunnel or other compression or entrapment neuropathies) leads to slowing of conduction and decrement of amplitude across the segment including the lesion. However, studies of the nerve distal to the lesion are fully normal. Studies of nerve segments proximal to the lesion reveal normal conduction velocity with an unchanging reduced action potential amplitude.

Quite dramatic nerve conduction findings are seen with a focal total neurapraxic lesion (eg, acute trauma or laceration). In this setting, although the nerve is fully normal below the lesion, electrical stimulation proximal to the lesion produces no response (similar to the patients' attempts to activate the muscle).

In radiculopathy, motor nerve conduction studies will ordinarily be normal. However, if the process leads to sufficient axonal loss, there may be slight slowing of conduction velocity in direct relation to the amount of loss of large fibers. In central nervous system disease, there will ordinarily be no change in motor nerve conduction unless there is involvement of anterior horn cells. In demyelinating neuropathies, there is slowing of conduction velocity and prolongation of distal motor latency.

Late responses — Studying the most proximal segments of nerves is difficult, as the nerves are deep and not easily accessible as they leave the spinal column. Nevertheless, it is useful to study the proximal segments of a nerve, since processes such as radiculopathies from disc protrusion and certain neuropathies such as Guillain-Barré affect this segment predominantly.

The so-called late responses, the H-reflex and the F-response, provide a relatively easy technique for the study of the proximal segments of nerves. These responses are produced in certain circumstances after an electrical stimulus to a peripheral nerve and are "late" with respect to the muscle response (the M-response) produced by the orthodromic volley of action potentials traveling to the muscle directly from the electrical stimulus.

H-reflex — The H-reflex is a monosynaptic reflex response that is similar in pathway to the tendon jerk. The electrical stimulus activates the Ia afferents (coming from the muscle spindles) and action potentials travel orthodromically to the spinal cord. In the cord, the Ia afferents make excitatory monosynaptic connections to the alpha motor neurons and a volley of action potentials is set up in the motor nerve that runs orthodromically the entire length of the nerve from the cell bodies to the muscle. Thus, action potentials travel through the proximal segment of the nerve twice in the production of the H-reflex, once in the sensory portion of the nerve and once in the motor portion.

Obtaining an H-reflex depends upon the ability to stimulate the Ia afferents without stimulating the motor axons. If motor axons are electrically stimulated, an action potential travels along the axon antidromically toward the spinal cord as well as orthodromically toward the muscle. The antidromic action potential collides (either in the proximal motor axon or cell body) with the developing H-reflex in that axon, thereby nullifying it. In routine clinical practice, it is possible to get this differential stimulation and produce H-reflexes only by stimulating the posterior tibial division of the sciatic nerve and recording from the triceps surae.

F-response — The F-response, or F-wave, has an advantage over the H-reflexes since it can be found in most muscles. The response is a manifestation of recurrent firing of an anterior horn cell after it has been invaded by an antidromic action potential. After a motor nerve is stimulated, an action potential runs antidromically as well as orthodromically; a small percentage of anterior horn cells that have been invaded antidromically produce an orthodromic action potential that is responsible for the F-response. Thus, to produce an F-response, an action potential must travel twice through the proximal segment of the motor nerve [17].

ELECTROMYOGRAPHY — The physiology of electromyography (EMG) is based upon the function of the motor unit [18]. A motor unit is composed of all the muscle fibers innervated by a single anterior horn cell. In most proximal limb muscles, there are hundreds of fibers in each motor unit. In healthy tissue, muscle fibers from the same unit are not clumped together, but are intermingled with fibers from other motor units. When a motor axon fires, each muscle fiber in its motor unit is activated in a constant time relationship to the other fibers in the unit.

EMG activity is ordinarily recorded with a needle placed into the muscle. Because the muscle fibers of a single motor unit are not packed close together, the EMG needle records from only about 10 fibers from each motor unit. The amplitude, duration, and configuration of the electrical activity recorded from a motor unit varies as the needle changes its orientation to the muscle fibers. Despite this variability, it is possible to specify a normal range for the amplitude, duration, and configuration of motor unit action potentials (MUAPs) for each muscle and each age. In relation to configuration, the number of phases above and below the baseline are counted. More than four phases is polyphasic and only a small percentage of polyphasic units in a muscle is considered normal.

When an EMG needle is placed in a normal muscle at rest, there is no electrical activity. With weak effort, first one and then several motor units are activated. At this low level of activation, it is possible to see the individual MUAPs and evaluate their parameters. With maximal effort, individual MUAPs cannot be discerned since a large number of units are brought into action. All that can be observed is a dense electrical pattern, called the interference pattern, which can be characterized by its density and peak-to-peak amplitude. The normal density would be either "full" if there are no gaps or "high mixed" if there are a few short gaps. Some patients are unable or unwilling to exert a maximal effort and the pattern will be less dense as a result. Hence, the degree of effort has to be taken into account when assessing the interference pattern. (show figure 19)

Electromyographic findings — Findings in EMG are best understood by describing results in different settings.

Acute partial axonotmestic injury — Acute partial axonotmestic injury describes a partial laceration of a nerve. The injured motor axons undergo Wallerian degeneration over the course of about five days. Muscle fibers previously innervated by those axons are left denervated. Within approximately 10 to 14 days, denervated muscle fibers develop spontaneous activity. These spontaneous muscle fiber action potentials are recorded by the EMG needle as fibrillations and positive sharp waves. Fibrillations and positive sharp waves are the same except for a slight difference in the particulars of the recording: both are simply small, diphasic potentials that begin with a positive phase (show figure 19).

The motor units that can be activated are normal. However, it is not possible to voluntarily activate the denervated muscle fibers. Thus, the interference pattern is less than full and the amplitude is decreased in proportion to the extent of the injury. Descriptive terminology for these patterns, in order of decreasing density, include high mixed, mixed, low mixed, and single unit.

After weeks to months, there will be collateral sprouting from surviving motor axons to innervate the denervated muscle fibers. Spontaneous activity will therefore cease. Since motor units contain more muscle fibers than normal, MUAPs will be long in duration, high in amplitude, and more complex in shape or polyphasic. The interference pattern may improve in density, but are likely to remain less than full (although the amplitude increases).

Complete neurapraxic injury — With complete neurapraxic injury (or complete laceration), no voluntarily initiated motor nerve action potentials can reach the muscle due to a focal demyelinating injury. However, the axons are intact so that the muscle fibers will not be denervated, and they will not fibrillate. EMG examination do not reveal spontaneous activity, MUAPs, or an interference pattern. This finding is not different from the first few days of a total axonotmestic injury. However, it is significantly different from total axonotmestic injury after the first few days, a time when denervated muscle fibers begin to fibrillate.

Myopathy — The simple model of myopathy is characterized by dropout of individual muscle fibers from their motor units. In active myopathies, especially polymyositis, there may be some segmental muscle necrosis. This process divides a muscle fiber into an innervated segment and an uninnervated segment. Since an uninnervated segment may fibrillate, some fibrillation and positive sharp waves may be present in active myopathies, although spontaneous activity is most commonly lacking (show figure 19). (See "Differential diagnosis of peripheral nerve and muscle disease" and see "Approach to the patient with muscle weakness").

NEUROMUSCULAR JUNCTION STUDIES — The most common abnormality at the neuromuscular junction is myasthenia gravis (MG) [18]. Other disorders include the Eaton-Lambert syndrome. (See "Clinical manifestations of myasthenia gravis").

Repetitive nerve stimulation study — A standard neurophysiological study for MG is the repetitive nerve stimulation study or Jolly test. The set-up is similar to that of a motor nerve conduction study. The nerve is stimulated and the muscle action potential is monitored. Normally, the action potential is the same each time; with MG, however, the action potential amplitude may decline on successive stimulations with rates of stimulation of 2 to 10 Hz. There are a variety of patterns that can be observed, but a good measure of this decline is a comparison of the amplitudes of the first and fourth potentials. A decline of more than 10 percent is considered pathologic (show figure 20).

In the inverse myasthenic syndrome or Eaton-Lambert syndrome, the Jolly test shows a progressive increase in the amplitude of the muscle action potential. In this setting, rates of 10 to 20 Hz are needed and a potentiation of 200 percent or more is required for the test to be considered pathologic. (See "Paraneoplastic syndromes affecting peripheral nerve and muscle").

Single fiber EMG — Single fiber EMG is sensitive for the diagnosis of MG [19]. In this test, a small needle is used that can detect the action potentials from a pair of muscle fibers of the same motor unit. The time between the firing of two muscle fibers from the same motor unit should be very stable. Due to the slow and uncertain function of the neuromuscular junction with MG, there may be an increase in the variability of the time between filling of these two muscle fibers. This is called increased jitter. (See "Diagnosis of myasthenia gravis").

Nystagmus is a rhythmic regular oscillation of the eyes. It may consist of alternating phases of a slow drift in one direction with a corrective quick "jerk" in the opposite direction, or of slow, sinusoidal, "pendular" oscillations to and fro. Jerk nystagmus is more common than pendular nystagmus. (See "Jerk nystagmus" and see "Pendular nystagmus").

Nystagmus can be continuous or paroxysmal, or evoked by certain maneuvers such as specific gaze or head positions. Nystagmus associated with some pathologies may only be seen transiently with hyperventilation or coughing and sneezing.

An overview of nystagmus, its treatment, and the vestibular physiology relevant to nystagmus and vertigo is presented here. The approach to vertigo is discussed separately. (See "Evaluation of vertigo").

TYPES OF NYSTAGMUS — As previously mentioned, the two major types of nystagmus are jerk nystagmus and pendular nystagmus.

Jerk nystagmus — Jerk nystagmus is subdivided by trajectory and the conditions under which it occurs (show table 1). Some forms are always present, even when the eyes are in the primary position. Nystagmus in the primary position is classified according to trajectory: downbeat; upbeat; horizontal; torsional; or mixed. The direction named is the direction of the fast phase. Other forms emerge only under specific conditions such as peripheral gaze (gaze-evoked) and certain head positions (positional). (See "Jerk nystagmus").

Most acquired jerk nystagmus are the result of an asymmetry in vestibular inputs, in either the central or peripheral nervous system. Understanding the diagnostic value of the trajectory of jerk nystagmus requires some knowledge of the physiology of the semicircular canals (see below).

Pendular nystagmus — Pendular nystagmus has a sinusoidal oscillation without fast phases. The waveform of pendular nystagmus may occur in any direction; it can be torsional, horizontal, vertical, or a combination of these, resulting in circular, oblique, or elliptical trajectories. It can be different in the two eyes, sometimes even monocular.

The diagnosis of pendular nystagmus is an exercise in pattern recognition, since pendular nystagmus is subdivided into a number of types recognized mainly by very specific trajectories (see "Pendular nystagmus"): Acquired pendular nystagmus Congenital nystagmus Pendular nystagmus with visual loss Specific variants: spasmus nutans; oculopalatal myoclonus; see-saw nystagmus; oculomasticatory myorhythmia (Whipple's)

BASIC CLINICAL VESTIBULAR PHYSIOLOGY

The vestibular sensory organs — There are two components to the peripheral vestibular system: The otolithic organs The semicircular canals

The otolithic organs are the saccule and the utricle, each of which has a plane of hair cells embedded in a heavy slab of material containing calcium. These slabs have a significant amount of inertia. Any linear acceleration of the head (eg, gravity) causes the head to move relative to the otoconia, resulting in a bending of the hair cells and a change in neuronal activity.

The semicircular canals detect angular accelerations (turns, or rotations) of the head. Each canal contains fluid and a gelatinous structure called the ampulla, in which hair cells are embedded. Rotation of the head causes the canal to move relative to the fluid within it, thereby generating a force against the ampulla that bends the hair cells. An increase or decrease from the baseline firing rate occurs depending upon the direction of rotation.

There are six canals, two per side, arranged along three cardinal planes. The two horizontal canals lie in one plane, not quite horizontal but tilted downward 30 degrees. The vertical (anterior and posterior) canals are all at approximately right angles to the horizontal plane, but positioned in two planes 45 degrees from the sagittal plane. Thus, each plane has a pair of canals, the anterior canal of one side paired with the posterior canal of the other side.

The vestibular system uses these pairs of canals in a "push-pull" reciprocal fashion. An angular movement of the head along one of these planes will maximally activate one canal and maximally inhibit its counterpart. Thus, the normal resting firing rate of the nerve from one canal increases while that of its counterpart decreases. The net change in both directions is summed algebraically by the brain to reveal the change in head position. This information is transmitted to the ocular motor nuclei, causing the eyes to move smoothly an equal and opposite amount to the head turn (the vestibuloocular reflex), enabling the eyes to remain stationary in space despite the head movement.

The directions that excite each canal are given by a simple rule: "a canal is excited by head motion towards the canal, in the appropriate plane." Thus, the right horizontal canal is excited by right head turns, the right posterior canal is excited by right head tilts or "posterior" head tilts (neck extension), and the left anterior canal is excited by left head tilts or "anterior" head tilts (neck flexion).

Vestibuloocular responses — The vestibuloocular reflex (or more correctly, the canaliculoocular reflex) keeps the line of sight stable in space while the head is moving. Thus, it uses information about head rotation to drive the eyes an equal amount in the opposite direction. To accomplish this, each semicircular canal has excitatory projections to a pair of extraocular muscles, one in each eye, and inhibitory projections to an antagonistic pair. The muscles maximally activated by a canal are approximately aligned with the plane of that canal.

To understand the effects of such activation requires knowledge of the direction of pull of each extraocular muscle. The horizontal recti are simple: the lateral recti abduct and the medial recti adduct the eye. The vertical recti and obliques are more complex: they all have primary, secondary, and tertiary actions upon the globe (show table 2). It is useful to think of the line of pull of these muscles as approximately aligned with the planes of the vertical semicircular canals. While the line of pull remains roughly constant, the effect of the pull upon the eye in part depends upon whether the eye is abducted or adducted at the time. Contraction of the superior rectus elevates the eye, especially when it is in the abducted position. When the eye is adducted, however, this strong primary action of the muscle is less and there is also more of an intorsional motion. Contraction of the inferior rectus leads to the opposite pattern: depression of the eye in abduction with some extorsion in adduction. Contraction of the superior oblique causes mainly intorsion of the eye, as it pulls the top of the eye towards the pulley at the anteromedial corner of the roof of the orbit. This is maximal in abduction. When the eye is adducted, however, the pull of the top of the eye toward the pulley now causes depression. Contraction of the inferior oblique, whose origin is at the anteromedial corner of the floor of the orbit, leads to the opposite pattern, extorsion of the eye in abduction, with elevation in adduction.

With these data it is easy to see which pair of muscles is activated (or inhibited) by each semicircular canal to keep the line of sight steady with head turns (show table 3).

Generation of nystagmus — Understanding the planes of action and ocular activations of the semicircular canals and their push-pull relationship with opposing canals on the other side is essential to understanding how jerk nystagmus is generated. With a lesion of the right horizontal canal, for example, there is a loss of tonic leftward vestibular input since there is always a baseline level of activity. The unopposed rightward "push" from the intact left horizontal canal creates a slow rightward drift of the eyes, the "slow phase," which is eventually counteracted by a leftward fast phase. Thus, loss of the right horizontal canal creates left-beating nystagmus.

Similarly, loss of the right anterior canal creates a nystagmus that is a mix of upbeat and counterclockwise torsional fast phases, and loss of the right posterior canal creates a downbeat/counterclockwise torsional nystagmus. If all three canals on the right side are lost, the upbeat and downbeat components cancel each other out, but the counterclockwise torsional effect add together and summate with the left-beating horizontal nystagmus to create a mixed horizontal-torsional nystagmus. This is the typical finding with "peripheral nystagmus," that is, nystagmus from acute unilateral damage to the labyrinth or vestibular nerve.

This also explains why pure vertical nystagmus is rarely if ever generated by peripheral vestibular disease. To obtain pure downbeat nystagmus, a lesion would have to affect the two posterior canals, summating the resulting loss of upward slow drifts, canceling the opposing torsional drifts, and sparing all other canals. The odds are against such a selective yet bilateral lesion of the labyrinth. The same holds for pure upbeat nystagmus. On the other hand, the central vestibular pathways in the brainstem are organized so that such deficits are possible.

Pure torsional nystagmus conceivably could result from a lesion that affected both the anterior and posterior canal on one side, sparing the horizontal canal. However, the regional blood supply and nerve subdivisions of the labyrinth segregate the anterior canal and horizontal canal from the posterior canal. Clinical studies suggest that vestibular neuronitis, for example, is a partial defect whose nystagmus suggests either a combination of horizontal and anterior canal dysfunction, or almost pure horizontal canal loss. Thus, partial lesions of the periphery are unlikely to cause pure torsional nystagmus. Pure horizontal nystagmus, however, could be either central or peripheral.

Some nystagmus is the result of hyperfunction of a canal rather than hypofunction of one labyrinth. The classic example is benign paroxysmal positional vertigo. Movement of debris in the posterior canal leads to abnormal activation, generating slow downward and torsional drifts, and therefore a mixed upbeat/torsional nystagmus. (See "Benign paroxysmal positional vertigo and nystagmus").

SYMPTOMS — Apart from vertigo, which is most prominent with acute unilateral peripheral vestibular disease, nystagmus can cause three main symptoms: Oscillopsia — Oscillopsia is a to-and-fro illusion of environmental motion. Depending upon the nystagmus, this may be continuous, intermittent, or gaze-evoked. This contrasts with the oscillopsia from bilateral vestibular hypofunction, which only occurs when the head is moving, as when walking or riding in a car. Blurred vision — Blurred vision occurs because the retinal image is smeared by stimulus motion. Abnormal head positions — Patients who find that their oscillopsia or blurred vision is least troublesome in certain gaze positions that minimize nystagmus may assume abnormal head positions.

DIFFERENTIAL DIAGNOSIS — The rhythmic nature and slow speeds of nystagmus distinguish it from a number of other abnormal involuntary eye movements. Saccadic intrusions are irregular bursts of rapid eye movements, almost always indicating a problem of the cerebellum. These movements include opsoclonus, ocular flutter, square wave jerks, and macrosaccadic oscillations. Voluntary nystagmus is a misnomer; it too is a type of saccadic intrusion, which most resembles ocular flutter. Ocular bobbing and its variants have slow drifts and fast phases but are much more irregular than nystagmus. They usually occur in comatose individuals. Oculogyric crises are irregular prolonged deviations of the eyes, usually up and lateral, that lack rhythmicity and slow phases. They are most frequently encountered with phenothiazine intoxication.

SYMPTOMATIC THERAPY — The general approach to nystagmus therapy involves three basic considerations [1,2]: What is the type of nystagmus? Are there treatable underlying causes? Consider intoxications, metabolic derangements, infections, and operable structural lesions. What is the potential benefit of symptomatic treatment?

The goals of improvement in the symptoms listed above (vertigo, oscillopsia, motion-degraded acuity, or head turn) should be clear prior to treatment. Patients without these symptoms do not require symptomatic treatment.

There are four main types of symptomatic nystagmus therapy: Medication Botulinum injections Prism lenses and optical solutions Surgery

Medication — Appropriate medication use depends upon the type of nystagmus. Proven drug therapies are only available for a few specific forms, such as baclofen for periodic alternating nystagmus [3] and gabapentin for pendular nystagmus [4]. (See "Pendular nystagmus"). There are unproven anecdotal reports for many others.

Botulinum injections — Botulinum injections into muscle or the intraconal space have been used to weaken the extraocular muscles and diminish the amplitude of nystagmus. These improve vision, oscillopsia, and nystagmus somewhat [5-7], but they also weaken normal eye movements. In some reports the most telling detail is that patients have declined to repeat the experience [6,8].

Prism lenses and optical solutions — Prism lenses and optical solutions may be useful in the following circumstances: If patients have a position of gaze in which nystagmus is minimal, prisms can be used so that the eyes are in this "null position" when the patient is trying to look directly ahead. Oscillopsia is thereby minimized for this most frequent gaze position, but gaze to the side will still bring on symptoms. If convergence dampens nystagmus, the patient can wear prisms that make the eyes more convergent while fixating in the distance [9,10]. High hyperopic glasses can be combined with high myopic contact lenses to create a situation in which the retinal image does not shift with eye movement [11,12]. This is useful for reading, but the field of view narrows to a tunnel, and the desirable effects of normal eye movements are impaired in that they do not shift or steady gaze normally. Thus, the system cannot be worn while walking around. Most patients find the device cumbersome. Contact lenses can dampen congenital nystagmus for unknown reasons [13].

Surgery — Most surgery tries to shift the attachment of the muscles so that the muscle tensions appropriate to a gaze position where nystagmus was minimal or absent now place the eyes in center gaze. In essence, this is similar to the strategy of prism therapy and, as with prisms, oscillopsia will still be present in lateral gaze. Surgery has been used mainly with congenital nystagmus [14-16], where it has been reported to improve visual acuity in some cases [14].