Somatosensory Evoked Potentials; Cerebral Potentials; Spinal Potentials; Nerve Stimulation; Sensory Lesions Muscle Nerve 21: 277–290, 1998

Home , Evoked potential

Key words: somatosensory evoked potentials; cerebral potentials; spinal potentials; nerve stimulation; sensory lesions Muscle Nerve 21: 277–290, 1998

AAEM MINIMONOGRAPH 19: SOMATOSENSORY EVOKED POTENTIALS

MICHAEL J. AMINOFF, MD, FRCP1 and ANDREW A. EISEN, MD2

1 Department of Neurology, University of California at San Francisco, San Francisco, California, USA 2 Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada

Received 11 August 1997; accepted 5 September 1997

More than 40 years have elapsed since the first so- sensory threshold. It is also possible to elicit SEPs matosensory evoked potential (SEP) was recorded using a variety of mechanical stimuli.63,105 This al- from humans,21 and 15 years have passed since the lows selective activation of specific sensory modali- clinical role of SEPs was reviewed as an American ties, but the SEPs elicited are often of small ampli- Association of Electrodiagnostic Medicine (AAEM) tude and may require many hundreds of responses minimonograph. At that time, interest in SEPs was to be averaged. This limits the clinical utility of me- reaching its zenith, and a vast literature had already chanically elicited SEPs. accumulated. The subsequent development of so- The SEP is greatly attenuated or abolished when phisticated imaging techniques has impacted on the the dorsal columns are selectively ablated in ani- role of SEPs in the clinical setting, making it timely mals,18 indicating that within the spinal cord the SEP to review this aspect as the end of the century is is mediated predominantly via these tracts. Con- approached. This minimonograph considers current versely, cord lesions that do not interrupt the dorsal concepts of the physiologic basis of the SEP, dis- columns are associated with a relatively normal SEP. cusses the different techniques available to elicit and Loss of posterior column function in humans is al- record it, and critically analyzes its clinical utility. most invariably accompanied by a grossly abnormal The peripheral electrical stimulus routinely used SEP.47 Some SEP components may, however, reflect to elicit an SEP activates predominantly—if not en- extralemniscal activity; they have been evoked in cats tirely—the large-diameter, fast-conducting group Ia 55 after selective dorsal column transection by stimuli muscle and group II cutaneous afferent fibers. Se- that are sufficient to excite small-diameter periph- lective intrafascicular stimulation has provided evi- eral fibers,76 and tourniquet-induced ischemia in hu- dence for a direct muscle afferent fiber (Ia) projec- mans abolishes short-latency before long-latency SEP tion to the human somatosensory cortex.43 However, components, suggesting that they are mediated by when a mixed nerve is stimulated, both group Ia different centrally conducting tracts.119 muscle afferents and cutaneous group II afferents Although in general the SEP is best recorded contribute to the resulting SEP.55 Its amplitude is over the somatosensory cortex, topographic map- almost maximal when the peripheral nerve action ping indicates that several of its components are potential is only 50% of its maximum.35 This trans- widely distributed over the scalp, and some are maxi- lates into a requisite stimulus intensity of about twice mally recorded outside the somatosensory cortex.112 Because the SEP monitors more than just the so- *Correspondence to: American Association of Electrodiagnostic Medi- matosensory pathways, abnormalities recorded in cine, 421 First Avenue S.W., Suite 300 East, Rochester, MN 55902, USA certain primary diseases of the motor system (such as CCC 0148-639X/98/030277-14 © 1998 American Association of Electrodiagnostic Medicine. Published amyotrophic lateral sclerosis) should not cause con- by John Wiley & Sons, Inc. cern.

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 277 STIMULATION TECHNIQUES Dermatomal Stimulation. Dermatomal stimulation Mixed Nerve Stimulation. Because electrical stimu- is even more segmentally specific than cutaneous lation of a mixed nerve initiates a relatively synchro- nerve stimulation, because cutaneous nerve stimula- nous volley that elicits a sizable SEP, it has become tion invariably activates fibers from more than one the standard for clinical use. The stimulus intensity dermatome. However, with dermatomal stimulation required to elicit an SEP of maximum amplitude the ascending volley is very desynchronized, and this need not be supramaximal. The ideal stimulus in- sometimes makes the SEP difficult to interpret. Der- duces a mixed nerve action potential that is just matomal stimulation has been used most often to greater than 50% of its maximum amplitude and assess function of the lumbosacral roots. For L5, the clinically produces a slight muscle twitch. Too high a medial side of the first metatarsophalangeal joint or stimulus is counterproductive, producing occlusion the dorsal surface of the foot between the first and of Ia impulse traffic by other converging afferent second toes is stimulated. For S1, the lateral side of impulses.43 The occurrence of occlusion may de- the fifth metatarsophalangeal joint is stimulated. pend on the limb that is stimulated.44 When occlu- Care must be taken to avoid stimulus spread to sion occurs, a percentage of the initial volley is inef- neighboring dermatomes, underlying muscle (which fective, and the SEP is accordingly lower in induces activity of Ia afferents), and digital cutane- amplitude. A stimulus of short duration (200–300 ous nerves. This can be achieved if the stimulus is µs) is popular, but longer duration (1000 µs) pulses kept at 2.5 × sensory threshold, which gives about of appropriately lower intensity may be appropriate 80% of the maximum amplitude. Normative data for because they preferentially activate the type Ia and II the L5 and S1 dermatomes are well established.4 afferents.113 A repetition rate of 5.1 Hz is conve- nient. Faster rates may be tolerable to the patient Motor Point Stimulation. A single muscle motor and do not alter the SEP until they exceed 15 Hz. point can be stimulated to elicit SEPs.31 This is Troublesome electrocardiographic artifact, achieved by using a monopolar needle electrode for particularly relevant with noncephalic referential re- stimulation. A long-duration (1.0 ms), low-intensity cordings of SEPs elicited from the legs, can be obvi- stimulus preferentially activates the Ia afferents that ated by triggering the stimulus off the electrocardio- exit in the same bundle as the alpha motor fibers at gram. In general, SEPs elicited by lower limb the motor point. This method allows stimulation of stimulation have the added advantage of testing the proximal (large) muscle Ia afferent input. functional integrity of much of the length of the spinal cord, an important consideration in suspected Paraspinal Stimulation. The paraspinal region at multiple sclerosis and other myelopathies that may sequential levels along the vertebral column can be not be detected by imaging techniques. stimulated to elicit SEPs.51 This method ‘‘obviates’’ peripheral input from long nerves in the limbs and Cutaneous Nerve Stimulation. Virtually any acces- more readily identifies lesions of the spinal cord. sible cutaneous nerve can be stimulated to elicit an The stimulus is applied simultaneously to both sides, SEP. With cutaneous nerve stimulation, however, the 2 cm lateral to the midline, at an intensity that in- potentials are smaller than those evoked by mixed duces a small visible muscle twitch. Potentials are nerve stimulation, and the small far-field compo- recorded over the scalp (Cz–Fz). The afferent volley nents are difficult to record.25,30 Use of cutaneous is primarily initiated in the cutaneous branches of nerve stimulation should be considered when it is the primary dorsal root rami, with some contribution necessary to: (1) assess the integrity of specific cuta- from the paraspinal Ia afferents. neous nerves, such as the lateral femoral cutaneous In normal subjects, the value for spinal cord con- nerve,116 which are not readily studied by conven- duction velocity between T12 and T1 is approxi- tional techniques; (2) measure peripheral sensory mately 64 m/s. However, it is better to refer to con- conduction when this is not otherwise possible be- duction time, which over the same segments is 5.4 ± cause the sensory nerve action potential (SNAP) is 1.6 ms.51 either absent or very small25,30; (3) evaluate isolated root function, because of the increased segmental RECORDING AND FILTERING OF SEPs specificity of cutaneous stimulation4,29; and (4) as- Surface or needle electrodes can be used for record- sess dubious patchy numbness for medicolegal rea- ing SEPs. There is no difference in the resultant SEP. sons, by stimulating homologous areas of ‘‘involved’’ The latter, although easily inserted into the scalp, and normal skin supplied through cutaneous termi- are not popular because of their higher impedance, nals.22 the discomfort of their insertion, and the theoretical

278 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 risk of infection. Recording montages are either ‘‘ce- may be recorded with bipolar cephalic derivations,37 phalic bipolar’’ or ‘‘referential.’’ In a cephalic bipo- but referential recording is required for proper lar montage both electrodes are placed on the head, identification. Far-field potentials are small in ampli- while in a referential montage the reference elec- tude, recorded with equal ease and amplitude over a trode is placed off the head. A cephalic bipolar mon- wide area of the scalp, and usually positive and tage is relatively noise-free and is satisfactory for rou- monophasic at the active electrode, reflecting a mov- tine clinical use. However, there is cancellation of ing front approaching the recording electrode. the small-amplitude, far-field potentials, which can Kimura and coworkers showed that under some cir- only be recorded with noncephalic references (e.g., cumstances far-field potentials may be biphasic and the opposite mastoid, shoulder, arm, hand, or knee, of either polarity.67 The posterior tibial near-field or the linked mastoids or earlobes). Recording with potential is also positive in polarity. both types of montage is advantageous (see Fig. 1). Small-amplitude components of the SEP are The number of channels (recording derivations) composed of both high and low frequencies, and used should be dictated by the specific reason for filtering can be problematic. Too wide a bandpass performing the SEP. For example, in field distribu- results in a ‘‘noisy’’ SEP, while a restrictive bandpass tion studies, 16 or more channels may be useful, attenuates either the high- or low-frequency compo- whereas one channel is sufficient when the SEP is nents, depending on the settings chosen. There is no being used to measure peripheral conduction veloc- ‘‘correct’’ filter setting—the choice is best related to ities. the particular task. For general purposes, a relatively With a bipolar cephalic derivation, near-field po- broad bandpass (10–2500 Hz) is suitable. Restrictive tentials, which are characteristically of negative po- filtering of between 150 and 300 Hz to 3000 Hz en- larity and of relatively large amplitude, are recorded. hances high-frequency, small-amplitude, near- and Their amplitude falls rapidly when the electrode is far-field components, but does so at the expense of moved only a short distance from the generator the low-frequency components.74,75 source. Certain small-amplitude far-field potentials Analog-restrictive filtering induces phase shift of

FIGURE 1. Normal SEP elicited by stimulation of the left median nerve and recorded over the Erb’s point region, fifth cervical spine, and scalp. EPi, ipsilateral Erb’s point; EPc, contralateral Erb’s point; CV5, fifth cervical spinous process; C3Ј and C4Ј, scalp placements halfway between C3 and P3, and C4 and P4, respectively. The bottom trace shows the Erb’s point potential, the second channel shows mainly the stationary cervical potential, the third channel shows the subcortical far-field potentials (especially P14 and N18), and the top channel shows the N20 potential.

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 279 components and may induce distortion of compo- by dermatomal—as opposed to nerve—stimulation nents. Digital filtering, designed for zero phase shift, are much more variable in amplitude. does not induce phase shift or distortion of compo- The morphology (shape) and dispersion of the nents37 and does not create ‘‘new’’ peaks. Digital SEP are difficult to quantitate, but both may be ab- filtering does however greatly enhance peaks that normal before or in the absence of latency prolon- are normally present. This is of particular promise gation or amplitude reduction. Computer-assisted for SEPs evoked by leg stimulation, when the small- methods for quantifying both dispersion and mor- amplitude components, usually difficult to recog- phology of the SEP have been described.36 nize, become easily identifiable. NEURAL GENERATORS OF THE SEP MEASUREMENT OF THE SEP It is generally accepted that the different compo- The latency, interpeak latency, amplitude, morphol- nents of the SEP predominantly reflect sequential ogy (presence or absence of components), and dis- activation of neural generators excited by the as- persion of the SEP can be measured, and side-to-side cending volley. This concept is appealing because it comparisons can be made. Latency is easily mea- provides a rational base for interpreting an absent or sured and standardized, whereas other characteris- abnormal component in relation to a specific ana- tics (e.g., morphology or dispersion) may be difficult tomic lesion. However, factors other than neural to assess. Latency varies with limb length. Interpeak generators (synapses in relay nuclei) are important transit (conduction) times are reliable parameters in the origin of some SEP components, especially the independent of limb length and usually indepen- small far-field potentials recorded using noncephalic dent of peripheral nerve disease. Central afferent references. In particular, the stationary far-field pathways do not mature at the same rate as periph- peaks, for example P9, P11, P13, and P14, evoked by eral pathways, and adult values for conduction veloc- median nerve stimulation reflect propagated volleys ity are not attained until 7 or 8 years of age.15 Aging of action potentials traveling in axons67 and can be is associated with prolongation of SEP latencies.24,62 recorded as the traveling volley approaches, but be- This is not simply a reflection of slowed peripheral fore it actually reaches, the active recording elec- conduction, because central conduction times are trode. This results from physical changes in the sur- also slowed significantly. rounding volume conductor,66,67 including the Absolute and interpeak latency are considered resistance or impedance of the volume conductor, abnormal when they are more than 3 SDs rather sites of axonal branching, and anatomic orientation than 2 or 2.5 SDs above the normal mean. Less sta- of the traveling impulse.66,67 tistical strictness is valid if the results are related to Of the far-field potentials in the median and age and height. In fact, however, latency is seldom tibial SEPs, it seems likely that relay nuclei are re- related to age and height in clinical laboratories. sponsible only for the generation of P13/14 (me- Absolute and interpeak latencies are easily measured dian) and P31 (tibial) components, with the other but frequently are normal in the face of obvious potentials probably reflecting electrophysical clinical impairment. Care is required to distinguish change in the surrounding volume conductor. between a delayed and an absent component; when The near-field N20 (median) and P38 (tibial) com- a component is absent, the subsequent component ponents are neurally generated, but it is likely that may be mistaken for the missing potential. they reflect multiple and even independent thalamo- Absolute amplitude of SEP components is vari- cortical projections.16,112,118 Certainly, from studies able, but a side-to-side difference exceeding 50% using digital filtering, it appears that there are ‘‘too is usually abnormal, providing the clinician can be many components’’ for available ‘‘generators.’’37 certain that the disease is unilateral. An excessive SEP components are designated by their polarity interside amplitude difference is nonspecific and in- and normal latency. The numerical latency values dicates either central conduction block or consider- ascribed to each component are summarized in able neuronal/axonal loss. However, complex facili- Table 1, which also indicates current views of the tatory and occlusive interactions leading to ‘‘central origin of specific, early latency, SEP components gain’’ can prevent amplitude reduction in the SEP, elicited by median and tibial nerve stimulation and masking axonal or neuronal loss.35,89 SEP amplitude recorded over the scalp.40,117 increases in the elderly,24 and ‘‘giant’’ potentials Little is known about the anatomic substrates of typify hereditary myoclonic epilepsy.101 SEPs evoked the middle (and longer) latency components of the

280 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 Table 1. Origin of far-field and near-field SEP components.

Generator Comments

Median nerve Far-field potentials P9 Just distal to brachial plexus Sometimes bilobed P9a and P9b (depending on arm position) P11 Dorsal root entry (presynaptic) P13/P14 Medial lemniscus (postsynaptic) P13 may reflect synaptic activity of dorsal horn interneurons N18 Between upper pons and midbrain N18 remains intact after thalamic lesions Near-field potential N20 Area 3b in the posterior bank of rolandic fissure Diagnostically the most useful peak to measure P22 Motor area 4 Increases with age P27 Parietal cortex N30 Supplementary motor area Decreases with age Tibial nerve Far-field potential P18 Sacral plexus Analogous to median P9 P31 Gracile nucleus Analogous to median P14 N34 Brain stem Analogous to median N18 Near-field potential P38/N38 Primary somatosensory cortex Analogous to median N20 Multiple cortical generators are involved

This reflects the authors’ view; differing opinions exist.

SEP. They exhibit considerable intertrial variability, level of the dorsal gray of the cervical cord. With a and their amplitude and latency are affected by cephalic reference these three components become sleep, level of consciousness, habituation, and cog- fused, but the latency difference between its peak nitive function. Nevertheless these components may and the N20 cortical SEP gives a ‘‘central conduction have some clinical relevance. Isolated abnormality of time’’ between the lower brain stem and cortex that the contralateral P40–N60 in the median SEP has measures about 5.5 ms. been described in multiple sclerosis and with deep Using a knee or iliac crest reference, SEPs elic- subcortical lesions involving thalamic nuclei other ited by tibial nerve stimulation can be recorded over than the primary sensory nuclei.117 the thoracolumbar spine. The components of the Long-latency components are distributed bilater- SEPs correspond to those described above for the ally, being of greatest amplitude at the vertex. These cervical SEPs elicited by median nerve stimulation. components relate to nonspecific thalamocortical Posterior tibial nerve stimulation at the ankle evokes projection systems involved in habituation, adapta- a traveling wave, N18, that can be recorded over the tion, and arousal. These components are of large lower lumbar spine immediately after the volley has amplitude, being between 10 and 40 µV, so that as passed the sacral plexus and is the scalp-recorded few as 50 averaged epochs are required to obtain a far-field P18 potential. This plexus-propagated volley measurable response. Unfortunately their recovery is the equivalent of the PPV recorded over the neck time takes several seconds, necessitating a stimulus with median nerve stimulation. A second distinct po- repetition rate that is lower than once every5sto tential, N22, can also be recorded over the lower avoid significant attenuation of the response. spine. It is a stationary peak and its latency remains constant. Its amplitude is maximal over the lower Spinal SEPs. Using a vertical and horizontal array thoracic spine (T10–L1). N22 probably represents of electrodes placed over the neck and referred to a postsynaptic activity in the lumbar gray matter gen- noncephalic reference, the cervical SEP elicited by erated in response to inputs from axon collaterals.40 median nerve stimulation consists of three distinct A third negative traveling wave, N24, whose latency components41: (1) the proximal plexus volley (PPV), shortens from caudal ro rostral thoracic cord, is evi- with a latency of about 10 ms; (2) the dorsal column dent over the lower thoracic cord. It reflects a prop- volley (DCV), reflecting the volley in the dorsal col- agated volley in the dorsal columns and is equivalent umn, with a latency of about 12 ms; and (3) the to the DCV potential recorded over the neck with cervical N13/P13, which is possibly generated at the median nerve stimulation. Spinal SEPs may not be

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 281 recordable in all normal subjects, and their absence Guillain–Barre´ syndrome when distal conduction is must be interpreted with caution. normal. The aim is to demonstrate conduction slow- ing in the proximal segments of peripheral nerves.10 SEPS IN DISORDERS OF THE PERIPHERAL AND In this regard, some physicians have found SEPs to CENTRAL NERVOUS SYSTEM be more sensitive than F waves,115 whereas others Although SEPs may reveal or localize a lesion involv- have found the converse.88,96 Aminoff and associ- ing the somatosensory pathways, they are simply an ates, in a study of 44 nerves in 15 patients with Guil- extension of the clinical examination and do not lain–Barre´ syndrome, found F-wave abnormalities indicate the underlying disease process. Normal more frequently than SEP abnormalities, especially findings do not exclude an organic basis for symp- when peripheral conduction was normal.88 The au- toms. Despite these limitations, SEP studies may be thors therefore believe that SEP studies are best re- helpful diagnostically and to determine the extent of served for occasions when both peripheral and F- pathologic involvement in different disorders. wave studies are normal in patients with suspected Guillain–Barre´ syndrome. Peripheral Nervous System. SEPs may have some role in the evaluation of Disorders of the Peripheral Nerves. SEPs may be patients with distal axonopathies, but this remains to helpful in evaluating the peripheral nervous system be established. when nerve conduction studies (NCSs) are inappli- Plexus Lesions. The prognosis of brachial plexus cable because of the proximal site or severity of pa- injuries influences clinical outcome; a postgangli- thology. Peipheral sensory conduction velocity can onic lesion—unlike root avulsion—sometimes re- be measured by SEPs with comparable results to quires operative treatment and may be followed by those from conventional NCSs. The nerve is stimu- recovery. Conventional NCSs can suggest the site of lated at two or more sites and the responses are re- injury. Preserved SNAPs despite clinical sensory loss corded over the scalp. Scalp-recorded SEPs may be indicate a preganglionic lesion. Recording over the absent or delayed in the presence of polyneuropa- scalp and spine of SEPs following stimulation of up- thies and mononeuropathies.9,47 Scalp-recorded per limb nerves has been suggested by some physi- SEPs may be important for identifying peripheral cians to improve the electrophysiologic evaluation of nerve lesions when SNAPs are unrecordable periph- plexus lesions, based on the belief that N13 attenu- erally. Parry and Aminoff recorded robust SEPs on ation reflects the total damage, whereas N9 attenu- stimulation of each of 15 nerves from which SNAPs 61 were absent or grossly attenuated as a result of ac- ation reflects the extent of postganglionic damage. quired demyelinating peripheral neuropathies.89 In Jones and coworkers compared the preoperative 11 instances there was slowed afferent conduction electrophysiologic findings with the site of the lesion at surgery in 16 patients with unilateral traction le- velocity (determined from the SEPs), but in four 61 nerves conduction velocity was within the normal sions of the brachial plexus. The electrophysi- range despite slowed motor conduction. The inabil- ologic findings correctly localized the lesion in only ity to record peripheral SNAPs in such cases must 8 instances; in 3, the findings suggested a purely reflect pathologic dispersal of impulses, in which postganglionic lesion in roots that were subsequently context the preservation of cortically generated SEPs found to be ruptured preganglionically. Moreover, a probably results from reorganization and synchroni- discrepancy sometimes occurred between the SEP zation of afferent volleys at different synaptic lev- and peripheral SNAP findings, perhaps because of els.25,31,35,89 If this is the case, the responses of a few recoil of an avulsed root from the spinal cord with its normally conducting surviving axons must account ganglion attached; this would affect the size and for the misleadingly normal values for conduction shape of the N9 potential recorded from about the velocity sometimes obtained by the SEP technique.89 clavicle, but not the distal SNAP.61 Focal nerve lesions can be diagnosed by scalp- The segmental level of proximal plexus lesions recorded SEPs,23 but this is only helpful when the has been delineated by recording SEPs to stimula- lesion is too proximal to be recognized by conven- tion of the median, radial, and ulnar nerves at the tional techniques, such as in meralgia pares- wrist,109 or the musculocutaneous nerve proximal to thetica.108 Some physicians have also used the SEP the wrist.107 With multiple root avulsions, there may technique to follow recovery after peripheral nerve be no spinal or scalp-recorded SEPs; with C7 root transection injuries—the SEP may be present when involvement, responses to radial nerve (but not me- peripheral SNAPs cannot be elicited. dian and ulnar nerve) stimulation may be abnor- The SEP findings have been used to diagnose mal.109 Cervical and cortical SEPs from musculocu-

282 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 taneous nerve stimulation are especially likely to be tion,48 but the clinical and electrophysiologic find- attenuated or absent with C5 and C6 root lesions.107 ings were not detailed, and interpretation is further A preganglionic lesion may be partially or com- complicated because diagnostic criteria for the SEP pletely obscured electrophysiologically by coexisting abnormalities were not provided. postganglionic pathology, thereby limiting the utility The published evidence is thus conflicting con- of SEPs for evaluating plexus lesions. Moreover, the cerning the role of SEPs in the evaluation of sus- information derived from SEP studies concerning pected thoracic outlet syndrome, but in the authors’ the site and extent of plexus injury can usually be view, SEPs are of limited utility in diagnosing the obtained by needle electromyography (EMG). Nev- neurogenic disorder and of no established value in ertheless, the finding of attenuated (but present) diagnosing the nonneurogenic variety. peripheral SNAPs and absent spine and scalp SEPs is Cervical Spondylotic Myeloradiculopathy. SEPs do important in suggesting that multiple root avulsions not distinguish spondylosis from other cervical le- have occurred in addition to more peripheral lesions. Patients with pain and paresthesias but no sions.6 neurologic signs frequently have normal median or At operation, the appearance of the brachial ulnar SEPs, whereas those with objective clinical defi- plexus may be misleading, especially when disrupted cits may have abnormal SEPs, with delayed or lost nerve fascicles appear intact. Intraoperative stimula- components, regardless of whether there is a my- tion of specific roots, with recording of the responses elopathy.39,45 The nature of the SEP abnormality from the contralateral scalp, is therefore helpful in does not indicate either the severity or long-term indicating whether there is functional continuity prognosis of the neurologic disorder, and some pa- with the cord.70,106 Similarly, when ruptured nerves tients with a spondylotic myelopathy have normal are to be repaired by grafting, cortical responses to SEPs.39,45 Thus the findings are no better than care- stimulation of the proximal stump indicate that a ful clinical examination in determining severity and second more rostral lesion is unlikely,70 and an ab- prognosis of cervical spondylosis, and do not help in sent response suggests a poor outcome. the selection of patients for surgery. Thoracic Outlet Syndrome. Characteristic electro- It has been suggested that fibers ascending from physiologic changes may occur in patients with neu- the lumbosacral region are more likely to be affected rogenic thoracic outlet syndrome. This is an uncom- by cervical compression than fibers arising from the mon disorder. In many patients suspected to have upper limbs.85 It is not yet established, however, neurogenic thoracic outlet syndrome, there is no whether recording sural SEPs in patients without a clinical or electrophysiologic abnormality in conven- myelopathy helps to indicate those at risk of a clini- tional studies. The SEP findings in such patients are cal cord deficit. variable. Both the authors’ findings,6 and those re- Epidurally recorded SEPs have been recorded in ported from the Mayo Clinic,120 indicate that SEP patients with cervical spondylosis but should not be studies are of limited value for the diagnosis of the used routinely for evaluating patients or in selecting neurogenic variety of thoracic outlet syndrome. In patients for surgery. some patients there may be abnormalities of the ul- Radiculopathy. There are reports that peroneal nar-derived SEP, such as an absent or markedly at- or tibial SEPs are abnormal in many patients with an tenuated N13 response despite a relatively normal isolated lumbosacral radiculopathy, but this is sur- N9 peak, or a small delayed N9 peak with or without prising, because stimulation of a multisegmental abnormalities of the N13 or N20 or prolonged inter- nerve should not lead to an SEP abnormality in the peak intervals. In one instance, the authors found presence of an isolated root lesion. Aminoff and col- the only abnormality was marked attenuation of the leagues have found normal peroneal SEPs in all of N20 component.6 their patients with a compressive L5 or S1 root le- In patients with the nonneurogenic syndrome, sion.4 the authors and others have found normal median, Eisen and coworkers stimulated cutaneous nerves ulnar, or radial SEPs.6,68,120 However, in one study, to improve the segmental specificity and recorded 12 of 18 patients, most of whom had no clinical defi- the SEPs in patients with myelographically proven cit, were found to have abnormal ulnar SEPs, with cervical or lumbosacral radiculopathies.32 Among 28 small N9 potentials despite preserved distal ulnar patients, only 16 (57%) had abnormal SEPs, particu- SNAPs, a finding difficult to explain on pathophysi- larly amplitude reductions and abnormal morphol- ologic grounds.59 In another study, 13 of 19 patients ogy; a prolongation in latency was uncommon. with suspected thoracic outlet syndrome had abnor- Others subsequently reported that SEPs elicited mal SEPs evoked by median or ulnar nerve stimula- by cutaneous nerve stimulation were more reward-

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 283 ing in detecting root involvement.91 Among 27 pa- ate patients with lumbar spinal stenosis, and may be tients with low back pain, unilateral radicular symp- more helpful in this context than when used to toms, and abnormal lumbosacral computerized evaluate isolated compressive root lesions (W.C. Sto- tomography (CT) scans, SEP abnormalities were lov, personal communication). A detailed account of found in 21 patients; in 15 patients there was no this work and a comparison of the yield of dermato- associated clinical deficit or other electrophysiologic mal SEPs and needle EMG, however, has yet to be abnormality. This suggested that the SEP was useful published. for detecting subclinical root pathology. More re- cently, however, Seyal and associates found with the Central Nervous System. The SEP findings may de- same technique that scalp-recorded SEPs were ab- tect and localize central somatosensory lesions but normal in only 20%—and lumbar responses in are not pathognomonic of specific diseases.1 50%—of patients with lumbosacral radiculopathies Detection of Lesions in Central Somatosensory Path- and accompanying radiologic abnormalities.100 ways. Multiple Sclerosis. The presence of SEP abnor- Thus, the preponderance of studies casts doubt on malities may reveal subclinical somatosensory lesions any role for SEPs elicited by cutaneous nerve stimu- and thereby establish that multiple lesions are pre- lation in the diagnosis or prognosis of radiculopa- sent. An SEP abnormality may also indicate that thies. vague sensory complaints have an organic basis, es- SEPs can be elicited by dermatomal stimulation pecially when the findings on clinical examination in the L5 or S1 territory.4,5,98 Scarff and colleagues are equivocal. reported abnormalities in 92% of patients with sur- The likelihood of finding an SEP abnormality is gically verified L5 or S1 root compression, but their greater in patients with definite multiple sclerosis criteria of abnormality were apparently derived arbi- (MS) than in patients with possible MS. The inci- trarily.98 Aminoff and associates studied 32 normal dence of abnormalities in patients with definite MS subjects and 19 patients with unilateral lumbosacral is about 80% regardless of whether there is any clini- radiculopathy, and found that in only 5 patients was cal sensory disturbance,46,77,102,111 but among pa- the lesion localized correctly by this technique.4 tients with possible MS the incidence of subclinical When the SEP was abnormal there was usually loss or SEP abnormalities is only about 25–35%. The yield is marked attenuation of the response; latency abnor- highest from SEPs elicited by stimulation of a nerve malities were uncommon. Needle EMG was gener- in the legs.1 An SEP abnormality is more likely if ally the most useful electrophysiologic means of there is a clinical sensory disturbance in the stimu- evaluating root lesions, but dermatomal SEPs some- lated limb, although the electrophysiologic finding times revealed unsuspected root dysfunction or ab- then provides little further information than ob- normalities when other electrophysiologic studies tained by clinical examination. SEP abnormalities were normal.5 are also more common when there are pyramidal Katifi and Sedgwick claimed that with dermato- signs in the stimulated limb20 or legs.103 In patients mal SEP studies they predicted correctly the pres- with a pyramidal deficit, however, an SEP abnormal- ence of root compression in 19 of 20 patients,64 but ity may not reflect a separate pathologic lesion but they used generous criteria for abnormality and merely an extensive lesion involving both motor and made multiple statistical comparisons of the data sensory pathways. from their patients, most of whom had extensive dis- Among the different evoked potential techniques ease.3 Thus, they correctly identified 25 of 34 roots for detecting subclinical involvement of afferent that were involved at surgery (their “gold” standard), pathways in patients with suspected MS, the highest but incorrectly identified another 12 roots that were yield is with SEPs; brain stem auditory evoked poten- not compressed.3 In fact, they accurately and com- tials (BAEPs) are the least useful.52,65,93,111 The basis pletely predicted on electrophysiologic grounds the of such differences is unknown but may reflect the operative findings in only 4 of 20 patients, a yield length and extent of the tracts being tested or dif- similar to that of Aminoff and associates.4,5 A more ferential susceptibility to pathologic involvement. recent study by Dumitru and Dreyfuss also suggested Abnormalities of the SEP may include delay or that dermatomal SEPs are of limited diagnostic util- absence of various components or an altered re- ity in patients with suspected unilateral L5 or S1 ra- sponse-morphology. In the median SEP, a common diculopathy, because they do not have both a high abnormality is loss of the cervical (N13) component sensitivity and a high specificity.28 They appear to be with preservation of the later components. Thus, in less accurate and sensitive than EMG and anatomic one study the cervical response was abnormal in 41% (imaging) studies.95 of patients with suspected MS, and the yield in- Dermatomal SEPs have also been used to evalu- creased to 50% when the cortical response was con-

284 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 sidered as well.38 The corresponding number for pa- may have MRI abnormalities resembling those in tients with definite MS was 87% and 73%, MS, so the MRI findings alone cannot establish the respectively. The cervical SEP abnormalities encoun- diagnosis. Further, evoked potential studies, particu- tered most often were absence, attenuation, and in- larly if used selectively in individual cases, are less creased dispersion of the response; only very occa- expensive than MRI. Finally, the electrophysiologic sionally was there an increase in latency. Central findings may clarify the underlying pathophysiology. conduction time, as calculated from the interpeak In certain patients with abnormal tibial SEPs, for latency of various components of the SEP, com- example, the P38 (P40) and succeeding negativity monly shows significant side-to-side difference in pa- are absent, and the first component recorded on the tients with MS.33 In normal subjects, there is no sig- scalp is a positive wave with a latency closely approxi- nificant change in the median SEP when body mating that of the P50, suggesting that conduction temperature is raised by 1°C, apart from a slight re- block (rather than delay) has occurred.34 duction in latency resulting from increased periph- SEPs may be abnormal in various disorders of the eral conduction velocity.79 Among patients with MS, central nervous system, other than MS, that inter- by contrast, an initially normal cervical response may rupt the somatosensory pathways. They are normal, become abnormal, or—less commonly—an initially however, when only spinothalamic function is im- abnormal response may normalize. In many pa- paired. tients, especially when the N13 is already abnormal, Spinal cord dysfunction. Patients with spinal cord it becomes disorganized and attenuated.79 tumors or malformations or with spinal arteriovenous Subclinical SEP abnormalities are not diagnostic malformations may have abnormal SEPs if the pos- of MS. Such SEP abnormalities may also be found in terior columns are involved.73 In patients with re- other disorders such as vitamin B12 and vitamin E paired spinal dysgraphic lesions, serial SEP studies deficiency states,42,69 hereditary spinocerebellar de- have sometimes been used to facilitate early detec- generations,87,90 and hereditary spastic paraple- tion of clinically significant retethering, but the find- gia.110 The electrophysiologic findings, therefore, ings do not correlate well with clinical status and are must always be interpreted in relation to the clinical of questionable utility.71 context. An upper cervical myelopathy occurs in achondro- It is not clear whether evoked potential tech- plasia due to a small foramen magnum, and may lead niques are useful for monitoring disease progression to abnormal median or peroneal SEPs.84 In one and evaluating experimental therapies for MS. Ami- study, SEPs were abnormal in all patients with neu- noff and associates found a poor correlation be- rologic symptoms or signs and in 44% of patients tween clinical and electrophysiologic changes over without neurologic dysfunction. In several of these time, as well as an excessive variability between test patients, CT scans showed significant stenosis at the sessions of the response to stimulation of a clinically foramen magnum. Thus SEPs may have some use as involved afferent pathway in patients with definite a monitor of cord function in achondroplasia. MS and a stable clinical deficit.2 Similar findings SEPs have also been used to monitor the effects have been reported by others for SEPs.72,80 Changes of radiation on the spinal cord. The findings have in previously abnormal evoked potentials do not nec- suggested that subclinical cord dysfunction may fol- essarily reflect the site of new lesions, and the elec- low radiotherapy for lung cancer even at conven- trophysiologic changes occurring during clinical ex- tional dose schedules.26 acerbations may be in pathways uninvolved clinically A variety of SEP abnormalities has been de- by the relapse. scribed in Friedreich’s ataxia, including mild latency Thus SEPs may be diagnostically helpful in pa- prolongation, and broadening or fractionation of tients with MS by establishing the presence of mul- the cortical response.78 Nuwer and colleagues found tiple lesions when this is clinically inapparent. This in Friedreich’s ataxia that SEPs elicited from upper- multiplicity of lesions is not, however, pathogno- limb nerves generally have delayed N20 peaks, but monic of MS and must be considered together with that the Erb’s point potential is either normal or any other clinical and laboratory findings. Any role only minimally delayed when present.87 In some for SEPs in following the course of MS or its re- cases the Erb’s point potential may be markedly at- sponse to treatment appears limited. tenuated.60 Evoked potential techniques and magnetic reso- In patients with hereditary spastic paraplegia or he- nance imaging (MRI) are probably equally sensitive reditary cerebellar ataxia, there may be loss of spinal in detecting lesions in patients with definite MS, but or cortical components or marked cortical delay, MRI is superior when the diagnosis is not definite.19 even when there is no clinical sensory loss.110 In However, patients without clinical evidence of MS these disorders and Friedreich’s ataxia, there may

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 285 also be abnormality of the visual and auditory evoked rietal N20–P27–P45 complex but not the prerolan- potentials.90 In olivopontocerebellar atrophy, SEPs dic P22–N30 complex, indicating that the N20 and may be normal,87 attenuated, or absent.56 P22 components have separate generators.81 Small SEPs may be abnormal in patients infected with postcentral lesions causing astereognosis (with oth- the human immunodeficiency virus (HIV). Patients with erwise preserved sensation) reduced or abolished acquired immune deficiency syndrome (AIDS) may the N20 and P27 components without affecting the have delayed cortical responses to tibial nerve stimu- P22–N30 complex. lation, suggesting slowed spinal conduction. In non- Recent work suggests that the electrophysiologic AIDS patients, spinal latency and conduction time findings are no better than a detailed neurologic from gluteal crest to T12 tend to increase over examination in predicting outcome from stroke,49 time.104 Again, others have reported that in some despite earlier suggestions to the contrary. patients with subclinical HIV infection, SEPs are ab- SEPs as a Guide to Prognosis. Coma. In 36 coma- normal and may suggest a conduction defect in cen- tose patients with preservation of some brain stem tral somatosensory pathways.13 Whether SEP studies function, Goldie and colleagues found the Erb’s have any clinical or prognostic relevance in this con- point response and the N13/N14 cervical response text is not established. In human T-cell lymphotropic to median nerve stimulation were preserved.50 Pa- virus (HTLV)-I-associated myelopathy/tropical spas- tients with bilateral loss of the N20–P22 response tic paraparesis, lower-limb SEPs may provide impor- either died or remained in a vegetative state. Unilat- tant information; central sensory conduction time eral loss of the N20–P22 responses was associated correlates with disability and may reveal subclinical with a variable clinical outcome: death, a persistent lesions of afferent pathways.83 Lower-limb SEPs may vegetative state, or recovery of a functional level. be unrecordable in patients with Pott’s paraplegia.82 Among patients with bilaterally preserved N20–P22 Brain stem lesions. SEPs are normal in Wallen- responses, 56% became functional, 25% remained berg’s syndrome,86 but are usually abnormal when in a vegetative state, and 19% died. Loss of the cor- the medial lemniscus is involved. They may be ab- tical N20–P22 responses therefore suggests a poor normal in the ‘‘locked-in’’ syndrome due to pontine prognosis. infarction.86 The SEP findings do not distinguish re- Among patients in posttraumatic coma, there is liably between intra- and extra-axial lesions of the an association between the degree of abnormality in brain stem.17 the median SEP and clinical outcome. Hume and Thalamic lesions. The SEP findings in patients Cant examined the scalp-recorded SEP findings with thalamic pathology tend to parallel the clinical soon after onset of posttraumatic coma.58 Bilateral findings, usually but not invariably being abnormal loss of median SEPs was believed to suggest a fatal when there are significant clinical deficits.57 Lateral outcome, and consistently asymmetric responses (in lesions affect particularly the tibial SEP, and more amplitude and latency) or a unilaterally absent SEP mesial lesions affect the median SEP.12 The SEP is were regarded as indicating the probable develop- sometimes relatively preserved with slowly growing ment of a severe residual deficit. The SEP findings noninfiltrating extrinsic tumors involving the thala- predicted the outcome corfectly in 38 of the 49 pa- mus, whereas extrinsic mass lesions compressing the tients studied in the first 84 h after coma onset. How- thalamus acutely may abolish both median and tibial ever, the SEP findings sometimes changed when re- SEPs. The SEP is generally markedly abnormal in corded serially, and the outcome did not correlate patients with intrinsic thalamic tumors and a clinical well with the SEP results in every test period. For sensory deficit.12 example, 44 patients ultimately did badly (i.e., died Hemispheric lesions. In patients with predomi- or were left with a severe deficit), and 10 of them nantly unilateral cerebral lesions, there is a correla- sometimes had normal SEP findings, while 3 had tion between the clinical and electrophysiologic normal SEPs on every recording. Similarly, among findings.47 SEPs are often normal in patients with no the 31 patients who did well, 15 had abnormal SEPs sensory deficit and abolished in those with moderate on at least one occasion. Thus the SEP findings may or severe cortical sensory loss. In some patients, how- not be as useful as originally hoped in providing a ever, there is a discrepancy between the clinical and satisfactory prognostic guide in individual cases. SEP findings—SEPs are normal despite cortical sen- Among patients in anoxic–ischemic coma, the sory deficits or absent despite preserved sensation. findings in one study revealed that none of 30 pa- Mauguiere and coworkers found that complete tients with absent cortical SEPs recovered cognition. parietal lesions produced contralateral hemianesthe- The presence of normal SEPs did not reliably predict sia without pyramidal signs, and eliminated the pa- recovery, however. Some patients with normal corti-

286 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 cal SEPs several hours after cardiac arrest recovered have found the SEP elicited from both upper and cognition, whereas others did not.11 The SEP find- lower limbs to be abnormal in patients with bulbo- ings corresponded to some extent with brain stem spinal neuronopathy (Kennedy’s syndrome).92 SEP function. One or several brain stem reflexes were studies have suggested that in diabetics, central affer- absent on initial examination in a majority of pa- ent transmission—as well as peripheral nerve con- tients with lost cortical SEPs but in only 20% of those duction—may be affected.14 Median SEPs may be with preserved SEPs. In another study, involving 60 abnormal in dystrophia myotonica, reflecting periph- patients in coma for more than 6 h after cardiac eral or central conduction delays in about one third arrest, the association of a Glasgow coma score of of cases.7 There seems to be no relationship, how- less than 8 at 48 h with abnormal or absent early ever, between the SEP findings and severity of the cortical components of the median SEP was highly clinical disorder. predictive of a bad outcome.8 Some patients have sensory symptoms but no ob- In locked-in syndrome, the SEP findings vary jective clinical deficits, suggesting that symptoms are considerably in different patients and are of little not organic in origin. In such a circumstance, an diagnostic help.54 abnormal SEP indicates that symptoms have an or- Brain Death. SEPs are more important than ganic basis. Normal SEPs and peripheral sensory BAEPs in evaluating patients with suspected brain NCSs may support suspicions that symptoms are not death, because all responses to auditory stimulation organic, but do not establish this with certainty. In- are often absent, possibly for technical reasons or deed, SEPs may be normal in patients with pure sen- from deafness.50 With median nerve stimulation, by sory stroke due, for example, to lacunar infarcts. contrast, most patients have cervical (N13/N14) but not later responses, indicating that input reaches the REFERENCES cord-medulla but does not generate cerebral activity more rostrally. This suggests an important role for 1. Aminoff MJ: Somatosensory evoked potentials in the evalu- SEPs in the evaluation of suspected brain death. The ation of the central nervous system. Neurol Clin 1988;6:809. 2. Aminoff MJ, Davis SL, Panitch HS: Serial evoked potential use of special recording derivations (midfrontal to studies in patients with definite multiple sclerosis: clinical median nasopharyngeal) for obtaining the median relevance. Arch Neurol 1984;41:1197. SEP may be helpful in distinguishing between coma 3. Aminoff MJ, Goodin DS: Dermatomal somatosensory evoked 114 potentials in lumbosacral root compression. J Neurol Neuro- and brain death, but this requires further study. surg Psychiatry 1988;51:740. Spinal Injuries. SEP changes may occur with spi- 4. Aminoff MJ, Goodin DS, Barbaro NM, Weinstein PR, Rosen- nal injury.27,97,99 Responses to stimulation of a nerve blum ML: Dermatomal somatosensory evoked potentials in unilateral lumbosacral radiculopathy. Ann Neurol 1985;17: in the legs may be normal, delayed, small, or lost, 171. depending on the extent and severity of the lesion 5. Aminoff MJ, Goodin DS, Parry GJ, Barbaro NM, Weinstein and the timing of the examination. PR, Rosenblum ML: Electrophysiologic evaluation of lumbosacral radiculopathies: electromyography, late responses and It may not be possible to record any cortical re- somatosensory evoked potentials. Neurology 1985;35:1514. sponse during the acute stage after incomplete trau- 6. Aminoff MJ, Olney RK, Parry GJ, Raskin NH: Relative utility matic cord lesions,99 so an absent SEP does not re- of different electrophysiologic techniques in the evaluation of brachial plexopathies. Neurology 1988;38:546. liably permit early distinction of complete from 7. Bartel PR, Lotz BP, Van der Meyden CH: Short-latency so- incomplete lesions. Preserved responses or their matosensory evoked potentials in dystrophia myotonica. J early return after injury, however, indicates an in- Neurol Neurosurg Psychiatry 1984;47:524. 8. Bassetti C, Bomio F, Mathis J, Hess CW: Early prognosis in complete lesion and therefore a better prognosis coma after cardiac arrest: a prospective clinical, electrophysi- than otherwise, although functional recovery may ological, and biochemical study of 60 patients. J Neurol Neu- still be poor.97 rosurg Psychiatry 1996;61:610. 9. Bergamini L, Bergamasco B, Fra L, Gandiglio G, Mombelli Use of SEPs to Prevent or Minimize Neurological- AM, Mutani R: Somatosensory evoked cortical potentials in Problems. SEPs have been used intraoperatively to subjects with peripheral nervous lesions. Electromyography monitor cord function, but their utility in this con- 1965;5:121. 10. Brown WF, Feasby TE: Sensory evoked potentials in Guillain- text is unclear and beyond the scope of the present Barre´ polyneuropathy. J Neurol Neurosurg Psychiatry 1984;47: minimonograph. 288. Defining the Extent of Neuropathologic Involvement 11. Brunko E, Zegers de Beyl D: Prognostic value of early cortical somatosensory evoked potentials after resuscitation from in Neurologic or Neuromuscular Disorders. SEP ab- cardiac arrest. Electroencephalogr Clin Neurophysiol 1987;66:15. normalities have been found in patients with amyo- 12. Chu N-S: Median and tibial somatosensory evoked poten- trophic lateral sclerosis,53,94 and this accords with pre- tials. Changes in short- and long-latency components in patients with lesions of the thalamus and thalamo-cortical ra- vious clinical and pathologic reports of sensory diations. J Neurol Sci 1986;76:199. involvement in this disorder. Polo and colleagues 13. Coats M, Jabbari B, Martin A, Salazar A: Serial evoked po-

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 287 tential studies in early human immune deficiency virus in- times in multiple sclerosis. Electroencephalogr Clin Neurophysiol fection. Neurology 1990;40(suppl 1):320. 1980;48:253. 14. Cracco J, Castells S, Mark E: Conduction velocity in periph- 34. Eisen A, Odusote K, Li D, Robertson W, Purvis S, Eisen K, eral nerve and spinal afferent pathways in juvenile diabetics. Paty D: Comparison of magnetic resonance imaging with Neurology 1980;30:370. somatosensory testing in MS suspects. Muscle Nerve 1987;10: 15. Cracco JB, Cracco RQ, Stolove R: Spinal evoked potential in 385. man: a maturational study. Electroencephalogr Clin Neurophysiol 35. Eisen A, Purves S, Hoirch M: Central nervous system ampli- 1979;46:58. fication: its potential in the diagnosis of early multiple scle- 16. Cracco RQ, Anziska BJ, Cracco JB, Vis GA, Rossini PM, Mac- rosis. Neurology 1982;32:359. cabee PJ: Short-latency somatosensory evoked potentials to 36. Eisen A, Roberts K, Lawrence P: Morphological measure- median and peroneal nerve stimulation: studies in normal ment of the SEP using a dynamic time warping algorithm. subjects and patients with neurological disease. Ann NY Acad Electroencephalogr Clin Neurophysiol 1986;65:136. Sci 1982;338:412. 37. Eisen A, Roberts K, Low M, Hoirch M, Lawrence P: Ques- 17. Csecsei GI, Klug N, Szekely G, Firsching RP, Christophis P: tions regarding the sequential neural generator theory of Multimodality electroneurophysiological findings in intra- the somatosensory evoked potential raised by digital filter- axial and extra-axial lesions of the brain stem. Acta Neurochir ing. Electroencephalogr Clin Neurophysiol 1984;59:388. 1995;137:48. 38. Eisen A, Stewart J, Nudleman K, Cosgrove JBR: Short-latency 18. Cusick JF, Myklebust JB, Larson SJ, Sances A Jr: Spinal cord somatosensory responses in multiple sclerosis. Neurology evaluation by cortical evoked responses. Arch Neurol 1979;36: 1979;29:827. 140. 39. EI Negamy E, Sedgwick EM: Delayed cervical somatosensory 19. Cutler JR, Aminoff MJ, Brant-Zawadzki M: Evaluation of pa- potentials in cervical spondylosis. J Neurol Neurosurg Psychiatry tients with multiple sclerosis by evoked potentials and mag- 1979;42:238. netic resonance imaging: a comparative study. Ann Neurol 40. Emerson RG: Anatomic and physiologic bases of posterior 1986;20:645. tibial nerve somatosensory evoked potentials. Neurol Clin 20. Davis SL, Aminoff MJ, Panitch HS: Clinical correlations of 1988;6:735. serial somatosensory evoked potentials in multiple sclerosis. 41. Emerson RG, Seyal M, Pedley TA: Somatosensory evoked Neurology 1985;35:359. potentials following median nerve stimulation: 1. The cervi- 21. Dawson GD: Cerebral responses to nerve stimulation in cal components. Brain 1984;107:169. man. Br Med Bull 1950;6:326. 42. Fine EJ, Hallett M: Neurophysiological study of subacute 22. Desmedt JE: Nature of focal lesion and location of skin combined degeneration. J Neurol Sci 1980;45:331. stimulus influence pattern and distribution of somatosenso- 43. Gandevia SC, Burke D, McKeon B: The projection of muscle ry evoked potentials (SEP): forensic and clinical uses. Neu- afferents from the hand to cerebral cortex in man. Brain rology 1979;29:585. 1984;107:1. 23. Desmedt JE: Somatosensory cerebral evoked potentials in 44. Gandevia SC, Burke D, McKeon BB: Convergence in the man, in Cobb WA (ed): Handbook of Electroencephalography somatosensory pathway between cutaneous afferents from and Clinical Neurophysiology. Amsterdam, Elsevier, 1971, vol 9, the index and middle fingers in man. Exp Brain Res 1983;50: p 55. 415. 24. Desmedt JE, Cheron G: Non-cephalic reference recording of 45. Ganes T: Somatosensory conduction times and peripheral, early somatosensory potentials to finger stimulation in adult cervical and cortical evoked potentials in patients with cer- or aging normal man: differentiation of widespread N18 and vical spondylosis. J Neurol Neurosurg Psychiatry 1980;43:683. contralateral N20 from the prerolandic P22 and N30 com- 46. Ganes T: Somatosensory evoked responses and central affer- ponents. Electroencephalogr Clin Neurophysiol 1981;52:553. ent conduction times in patients with multiple sclerosis. J 25. Desmedt JE, Noel P: Average cerebral evoked potentials in Neurol Neurosurg Psychiatry 1980;43:948. the evaluation of lesions of the sensory nerves and of the 47. Giblin DR: Somatosensory evoked potentials in healthy sub- central somatosensory pathway, in Desmedt JE (ed): New jects and in patients with lesions of the nervous system. Ann Developments in Electromyography and Clinical Neurophysiology. NY Acad Sci 1964;112:93. Basel, Karger, 1973, vol 2, p 352. 48. Glover JL, Worth RM, Bendick PJ, Hall PV, Markand OM: 26. Dorfman LJ, Donaldson SS, Gupta PR, Bosley TM: Electro- Evoked responses in the diagnosis of thoracic outlet syn- physiologic evidence of subclinical injury to the posterior drome. Surgery 1981;89:86. columns of the human spinal cord after therapeutic radia- 49. Goff PS, Karnaze DS, Fisher M: Assessment of median nerve tion. Cancer 1982;50:2815. somatosensory evoked potentials in cerebral ischemia. Stroke 27. Dorfman LJ, Perkash I, Bosley TM, Cummins KL: Use of cerebral evoked potentials to evaluate spinal somatosensory 1990;21:1167. function in patients with traumatic and surgical myelopa- 50. Goldie WD, Chiappa KH, Young RR, Brooks EB: Brainstem thies. J Neurosurg 1980;52:654. auditory and short-latency somatosensory evoked responses 28. Dumitru D, Dreyfuss P: Dermatomal/segmental somatosen- in brain death. Neurology 1981;31:248. sory evoked potential evaluation of L5/S1 unilateral/ 51. Goodridge A, Eisen A, Hoirch M: Paraspinal stimulation to unilevel radiculopathies. Muscle Nerve 1996;19:442. elicit somatosensory evoked potentials: an approach to 29. Eisen A: SEP in the evaluation of disorders of the peripheral physiological localization of spinal lesions. Electroencephalogr nervous system, in Cracco RQ, Bodis-Wollner I (eds): Fron- Clin Neurophysiol 1987;68:268. tiers of Clinical Neuroscience: Evoked Potentials. Baltimore, Wil- 52. Green JB, Price R, Woodbury SG: Short-latency somatosen- liams & Wilkins, 1986, p 409. sory evoked potentials in multiple sclerosis: comparison with 30. Eisen A: The somatosensory evoked potential. Can J Neurol auditory and visual evoked potentials. Arch Neurol 1980;37: Sci 1982;9:65. 630. 31. Eisen A: The use of somatosensory evoked potentials for the 53. Gregory R, Mills K, Donaghy M: Progressive sensory nerve evaluation of the peripheral nervous system. Neurol Clin dysfunction in amyotrophic lateral sclerosis: a prospective 1988;6:825. clinical and neurophysiological study. J Neurol 1993;240:309. 32. Eisen A, Hoirch M, Moll A: Evaluation of radiculopathies by 54. Gutling E, Isenmann S, Wichmann W: Electrophysiology in segmental stimulation and somatosensory evoked potentials. the locked-in syndrome. Neurology 1996;46:1092. Can J Neurol Sci 1983;10:178. 55. Halonen JP, Jones S, Shawkat F: Contribution of cutaneous 33. Eisen A, Odusote K: Central and peripheral conduction and muscle afferent fibres to cortical SEPs following median

288 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 and radial nerve stimulation in man. Electroencephalogr Clin evoked potentials in diagnosis of multiple sclerosis. Br Med J Neurophysiol 1988;71:331. 1976;2:25. 56. Hammond EJ, Wilder BJ: Evoked potentials in olivoponto- 78. Mastaglia FL, Black JL, Edis R, Collins DWK: The contribu- cerebellar atrophy. Arch Neurol 1983;40:366. tion of evoked potentials in the functional assessment of the 57. Hammond EJ, Wilder BJ, Ballinger WE: Electrophysiologic somatosensory pathway, in Tyrer JH, Eadie MJ (eds): Clinical recordings in a patient with a discrete unilateral thalamic and Experimental Neurology: Proceedings of the Australasian Asso- infarction. J Neurol Neurosurg Psychiatry 1982;45:640. ciation of Neurologists. Sydney, Adis Press, 1978, vol 15, p 279. 58. Hume AL, Cant BR: Central somatosensory conduction after 79. Matthews WB, Read DJ, Pountney E: Effect of raising body head injury. Ann Neurol 1981;10:411. temperature on visual and somatosensory evoked potentials 59. Jerrett SA, Cuzzone LJ, Pasternak BM: Thoracic outlet syn- in patients with multiple sclerosis. J Neurol Neurosurg Psychia- drome: electrophysiologic reappraisal. Arch Neurol 1984;41: try 1979;42:250. 960. 80. Matthews WB, Small DG: Serial recording of visual and so- 60. Jones SJ, Baraitser M, Halliday AM: Peripheral and central matosensory evoked potentials in multiple sclerosis. J Neurol somatosensory nerve conduction defects in Friedreich’s Sci 1979;40:11. ataxia. J Neurol Neurosurg Psychiatry 1980;43:495. 81. Mauguiere F, Desmedt JE, Courjon J: Astereognosis and dis- 61. Jones SJ, Wynn Parry CB, Landi A: Diagnosis of brachial sociated loss of frontal or parietal components of somato- plexus traction lesions by sensory nerve action potentials and sensory evoked potentials in hemispheric lesions. Brain somatosensory evoked potentials. Injury 1981;12:376. 1983;106:271. 62. Kakigi R: The effect of aging on somatosensory evoked po- 82. Misra UK, Kalita J: Somatosensory and motor evoked poten- tentials following stimulation of the posterior tibial nerve in tial changes in patients with Pott’s paraplegia. Spinal Cord man. Electroencephalogr Clin Neurophysiol 1987;68:277. 1996;34:272. 63. Kakigi R, Shibasaki H: Scalp topography of mechanically and 83. Moritoyo H, Arimura K, Arimura Y, Tokimura Y, Rosales R, electrically evoked somatosensory potentials in man. Electro- Osame M: Study of lower limb somatosensory evoked poten- encephalogr Clin Neurophysiol 1984;59:44. tials in 96 cases of HTLV-I-associated myelopathy/tropical 64. Katifi HA, Sedgwick EM: Evaluation of the dermatomal so- spastic paraparesis. J Neurol Sci 1996;138:78. matosensory evoked potential in the diagnosis of lumbo- 84. Nelson FW, Goldie WD, Hecht JT, Butler IJ, Scott CI: Short- sacral root compression. J Neurol Neurosurg Psychiatry 1987; latency somatosensory evoked potentials in the management 50:1204. of patients with achondroplasia. Neurology 1984;34:1053. 65. Khoshbin S, Hallett M: Multimodality evoked potentials and 85. Noel P, Desmedt JE: Cerebral and far-field somatosensory blink reflex in multiple sclerosis. Neurology 1981;31:138. evoked potentials in neurological disorders involving the 66. Kimura J, Mitsudome A, Beck DO, Yamada T, Dickins QS: cervical spinal cord, brainstem, thalamus and cortex, in Des- Field distribution of antidromically activated digital nerve medt JE (ed): Clinical Uses of Cerebral, Brainstem and Spinal potentials: model for far-field recording. Neurology 1983;33: Somatosensory Evoked Potentials. Basel, Karger, 1980, p 205. 1164. 86. Noel P, Desmedt JE: Somatosensory cerebral evoked poten- 67. Kimura J, Mitsudome A, Yamada T, Dickins QS: Stationary tials after vascular lesions of the brain-stem and diencepha- peaks from a moving source in far-field recording. Electroen- lon. Brain 1975;98:113. cephalogr Clin Neurophysiol 1984;58:351. 87. Nuwer MR, Perlman SL, Packwood JW, Kark RAP: Evoked 68. Komanetsky RM, Novak CB, Mackinon SE, Russo MH, Pad- potential abnormalities in the various inherited ataxias. Ann berg AM, Louis S: Somatosensory evoked potentials fail to Neurol 1983;13:20. diagnose thoracic outlet syndrome. J Hand Surg [Am] 1996; 88. Olney RK, Aminoff MJ: Electrodiagnostic features of the 21:662. Guillain-Barre´ syndrome: the relative sensitivity of different 69. Krumholz A, Weiss HD, Goldstein PJ, Harris K: Evoked re- techniques. Neurology 1990;40:471. sponses in vitamin B12 deficiency. Ann Neurol 1981;9:407. 89. Parry GJ, Aminoff MJ: Somatosensory evoked potentials in 70. Landi A, Copeland SA, Wynn Parry CB, Jones SJ: The role of chronic acquired demyelinating peripheral neuropathy. somatosensory evoked potentials and nerve conduction stud- Neurology 1987;37:313. ies in the surgical management of brachial plexus injuries. J 90. Pedersen L, Trojaborg W: Visual, auditory and somatosensory pathway involvement in hereditary cerebellar ataxia, Bone Joint Surg 1980;62-B:492. Friedreich’s ataxia and familial spastic paraplegia. Electroen- 71. Li V, Albright AL, Sclabassi R, Pang D: The role of somato- cephalogr Clin Neurophysiol 1981;52:283. sensory evoked potentials in the evaluation of spinal cord 91. Perlik S, Fisher MA, Patel DV, Slack C: On the usefulness of retethering. Pediatr Neurosurg 1996;24:126. somatosensory evoked responses for the evaluation of lower 72. Likosky W, Elmore RS: Exacerbation detection in multiple back pain. Arch Neurol 1986;43:907. sclerosis by clinical and evoked potential techniques: a pre- 92. Polo A, Teatini F, D’Anna S, Manganotti P, Salviati A, Dal- liminary report, in Courjon J, Mauguiere F, Revol M (eds): lapiccola B, Zanette G, Rizzuto N: Sensory involvement in Clinical Applications of Evoked Potentials in Neurology. New York, X-linked spino-bulbar muscular atrophy (Kennedy’s syn- Raven Press, 1982, p 535. drome): an electrophysiological study. J Neurol 1996;243:388. 73. Linden D, Berlit P: Spinal arteriovenous malformations: 93. Purves SJ, Low MD, Galloway J, Reeves B: A comparison of clinical and neurophysiological findings. J Neurol 1996;243:9. visual, brainstem auditory, and somatosensory evoked poten- 74. Lueders H, Andrish J, Gurd A, Weiker G, Klem G: Origin of tials in multiple sclerosis. Can J Neurol Sci 1981;8:15. far-field subcortical potentials evoked by stimulation of the 94. Radtke RA, Erwin A, Erwin CW: Abnormal sensory evoked posterior tibial nerve. Electroencephalogr Clin Neurophysiol potentials in amyotrophic lateral sclerosis. Neurology 1986;36: 1981;52:336. 796. 75. Maccabee PJ, Pinkhasov EI, Cracco RQ: Short latency so- 95. Rodriquez AA, Kanis L, Rodriquez AA, Lane D: Somatosen- matosensory evoked potentials to median nerve stimulation: sory evoked potentials from dermatomal stimulation as an effect of low frequency filter. Electroencephalogr Clin Neuro- indicator of L5 and S1 radiculopathy. Arch Phys Med Rehabil physiol 1983;55:34. 1987;68:366. 76. Martin HF, Katz S, Blackburn JG: Effects of spinal cord le- 96. Ropper AH, Chiappa KH: Evoked potentials in Guillain- sions on somatic evoked potentials altered by interactions Barre´ syndrome. Neurology 1986;36:587. between afferent inputs. Electroencephalogr Clin Neurophysiol 97. Rowed DW, McLean JAG, Tator CH: Somatosensory evoked 1980;50:186. potentials in acute spinal cord injury: prognostic value. Surg 77. Mastaglia FL, Black JL, Collins DWK: Visual and spinal Neurol 1978;9:203.

AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998 289 98. Scarff TB, Dallmann DE, Toleikis JR, Bunch WH: Dermato- somatosensory evoked potentials in hereditary spastic mal somatosensory evoked potentials in the diagnosis of paraplegia. J Neurol Neurosurg Psychiatry 1981;44:243. lumbar root entrapment. Surg Forum 1981;32:489. 111. Trojaborg W, Petersen E: Visual and somatosensory evoked 99. Sedgwick EM, El-Negamy E, Frankel H: Spinal-cord poten- cortical potentials in multiple sclerosis. J Neurol Neurosurg tials in traumatic paraplegia and quadriplegia. J Neurol Neu- Psychiatry 1979;42:323. rosurg Psychiatry 1980;43:823. 112. Tsuji S, Murai Y, Hashimoto M: Frontal distribution of early 100. Seyal M, Sandhu LS, Mack YP: Spinal segmental somatosen- cortical somatosensory evoked potentials to median nerve sory evoked potentials in lumbosacral radiculopathies. Neu- stimulation. Electroencephalogr Clin Neurophysiol 1988;71:273. rology 1989;39:801. 113. Veale JL, Mark RF, Rees S: Differential sensitivity of motor 101. Shibasaki H, Yamashita Y, Kuroiwa Y: Electroencephalo- and sensory fibres in human ulnar nerve. J Neurol Neurosurg graphic studies of myoclonus. Brain 1978;101:447. Psychiatry 1973;36:75. 102. Small DG, Beauchamp M, Matthews WB: Spinal evoked po- 114. Wagner W: Scalp, earlobe and nasopharyngeal recordings of tentials in multiple sclerosis. Electroencephalogr Clin Neuro- the median nerve somatosensory evoked P14 potentials in physiol 1977;42:141. coma and brain death. Detailed latency and amplitude 103. Small DG, Matthews WB, Small M: The cervical somatosen- analysis in 181 patients. Brain 1996;119:1507. sory evoked potential (SEP) in the diagnosis of multiple 115. Walsh JC, Yiannikas C, McLeod JG: Abnormalities of proxi- sclerosis. J Neurol Sci 1978;35:211. mal conduction in acute idiopathic polyneuritis: comparison 104. Smith T, Jakobsen J, Trojaborg W: Somatosensory evoked of short latency evoked potentials and F-waves. J Neurol Neu- potentials during human immunodeficiency virus (HIV) in- rosurg Psychiatry 1984;47:197. fection. Electroencephalogr Clin Neurophysiol 1990;75:S142. 105. Starr A, McKeon B, Skuse N, Burke D: Cerebral potentials 116. Wang J, Cohen LG, Hallett M: Scalp topography of somato- evoked by muscle stretch in man. Brain 1981;104:149. sensory evoked potentials following electrical stimulation of 106. Sugioka H, Tsuyama N, Hara T, Nagano A, Tachibana S, femoral nerve. Electroencephalogr Clin Neurophysiol 1989;74: Ochiai N: Investigation of brachial plexus injuries by intra- 112. operative cortical somatosensory evoked potentials. Arch Or- 117. Yamada T: The anatomic and physiologic bases of median thop Trauma Surg 1982;99:143. nerve somatosensory evoked potentials. Neurol Clin 1988;6: 107. Synek V: Somatosensory evoked potentials from musculocu- 705. taneous nerve in the diagnosis of brachial plexus injuries. J 118. Yamada T, Kayamori R, Kimura J, Beck DO: Topography of Neurol Sci 1983;61:443. somatosensory evoked potentials after stimulation of the me- 108. Synek VM: Assessing sensory involvement in lower limb dian nerve. Electroencephalogr Clin Neurophysiol 1984;59:29. nerve lesions using somatosensory evoked potential tech- 119. Yamada T, Muroga T, Kimura J: Tourniquet-induced isch- niques. Muscle Nerve 1985;8:511. emia and somatosensory evoked potentials. Neurology 1981; 109. Synek VM, Cowan JC: Somatosensory evoked potentials in 31:1524. patients with supraclavicular brachial plexus injuries. Neurol- 120. Yiannikas C, Walsh JC: Somatosensory evoked responses in ogy 1982;32:1347. the diagnosis of thoracic outlet syndrome. J Neurol Neurosurg 110. Thomas PK, Jefferys JGR, Smith IS, Loulakakis D: Spinal Psychiatry 1983;46:234.

290 AAEM Minimonograph 19: SEPs MUSCLE & NERVE March 1998