You are in: eMedicine Specialties > Neurology > Electroencephalography and Evoked Potentials Somatosensory Evoked Potentials: General PrinciplesArticle Last Updated: May 15, 2006AUTHOR AND EDITOR INFORMATIONSection 1 of 10
 
Author: Alan D Legatt, MD, PhD, Laboratory Director, Professor of Clinical Neurology, Department of Neurology, Montefiore Medical Center, Albert Einstein College of Medicine Alan D Legatt is a member of the following medical societies: Alpha Omega Alpha, American Academy of Neurology, American Clinical Neurophysiology Society, Eastern Association of Electroencephalographers, and New York Academy of Sciences Coauthor(s): Emad Soliman, MD, MSc, Consulting Staff, Department of Neurology, St John's Riverside Hospital Editors: Anthony M Murro, MD, Laboratory Director, Professor, Department of Neurology, Medical College of Georgia; Francisco Talavera, PharmD, PhD, Senior Pharmacy Editor, eMedicine; Norberto Alvarez, MD, Assistant Professor, Department of Neurology, Harvard Medical School; Consulting Staff, Department of Neurology, Boston Children's Hospital; Selim R Benbadis, MD, Professor, Director of Comprehensive Epilepsy Program, Departments of Neurology and Neurosurgery, University of South Florida School of Medicine, Tampa General Hospital; Nicholas Y Lorenzo, MD, Chief Editor, eMedicine Neurology; Consulting Staff, Neurology Specialists and Consultants Author and Editor Disclosure Synonyms and related keywords: SSEP, evoked potentials, somatosensory-evoked potentials, somatosensory evoked potential, somatosensory-evoked potential, evoked potential, impedances, montages, monophasic, nervous system, central nervous system, CNS, sensory stimuli, nerve, nerves, SSEPs, somatosensory pathways INTRODUCTIONSection 2 of 10
 
Evoked potentials are the electrical signals generated by the nervous system in response to sensory stimuli. Auditory, visual, and somatosensory stimuli are used commonly for clinical evoked potential studies. Somatosensory evoked potentials (SSEPs) consist of a series of waves that reflect sequential activation of neural structures along the somatosensory pathways. While SSEPs can be elicited by mechanical stimulation, clinical studies utilize electrical stimulation of peripheral nerves, which gives larger and more robust responses. The stimulation sites typically used for clinical diagnostic SSEP studies are the median nerve at the wrist, the common peroneal nerve at the knee, and/or the posterior tibial nerve at the ankle. Recording electrodes are placed over the scalp, the spine, and peripheral nerves proximal to the stimulation site. The dorsal column-lemniscal system is the major anatomical substrate of the SSEPs within the CNS. SSEPs are used for clinical diagnosis in patients with neurologic disease for prognostication in comatose patients and for intraoperative monitoring during surgeries that place parts of the somatosensory pathways at risk. Abnormal SSEPs can result from dysfunction at the level of the peripheral nerve, plexus, spinal root, spinal cord, brain stem, thalamocortical projections, or primary somatosensory cortex. Since individuals have multiple parallel afferent somatosensory pathways (eg, anterior spinothalamic tract and dorsal column tracts within the spinal cord), recordings of SSEPs can be normal even in patients with significant sensory deficits. SSEPs depend on the functional integrity of the rapidly conducting, large-diameter group IA muscle afferent fibers and group II cutaneous afferent fibers, which travel in the posterior column of the spinal cord. When a mixed peripheral nerve (with both sensory and motor components) is stimulated, both group IA muscle afferents and group II cutaneous afferents contribute to the resulting SSEP. Selective ablation of the dorsal column of the spinal cord abolishes the SSEPs generated rostral to the lesion. Diseases of the dorsal columns in which joint position sense and proprioception are impaired invariably are associated with abnormal SSEPs. The development of and easy access to sophisticated neuroradiologic imaging have had a great impact on the usage of SSEPs in clinical settings; fewer diagnostic SSEP studies are being performed now than in the pre-MRI era. Nevertheless, SSEPs are valuable as a diagnostic test in several clinical situations. Their role in the operating room has expanded, and interest remains high in SSEPs as research tools for unraveling of fundamental aspects of sensory physiology. An overview of the general principles of recording and interpretation of SSEPs is presented in this article. ELECTRICAL STIMULATION PARAMETERSSection 3 of 10
Stimulus location For recording median nerve SSEPs, the nerve is stimulated at the wrist. The anode is placed just proximal to the palmar crease, and the cathode is placed between the tendons of the palmaris longus muscle, 3 cm proximal to the anode. Ulnar nerve SSEPs are preferred to median nerve SSEPs for assessing the lower cervical spinal cord, especially during intraoperative monitoring when this part of the spinal cord is at risk, because the ulnar nerve originates from spinal roots C8-T1, whereas the median nerve originates from C6-T1. Ulnar nerve SSEPs will detect dorsal column damage at the C7-C8 level, but the median nerve SSEPs may persist, mediated by sensory fibers that enter the spinal cord at the C6 level. For recording posterior tibial nerve SSEPs, the nerve is stimulated at the ankle, with the cathode midway between the Achilles tendon and the medial malleolus and the anode 3 cm distal to the cathode. For recording peroneal nerve SSEPs, the common peroneal nerve is stimulated at the knee, with the cathode inferior to the leg crease just medial to the tendon of the biceps femoris muscle and the anode 3 cm distal to the cathode. In the lower limb, posterior tibial SSEPs are generally preferred because of the following:
However, peroneal nerve stimulation may be better in some cases because of the following:
Stimulus intensity The selected nerves are stimulated with monophasic square pulses, 100-300 microseconds in duration. Stimuli are delivered by using either a constant voltage or a constant current stimulator. The contact impedances of the stimulating electrodes should be kept low for the following reasons:
In the clinical setting, the stimulus intensity is set high enough to produce a consistent muscle twitch, which usually is tolerable to the patient. Because the patient is anesthetized during intraoperative SSEP monitoring, higher stimulus intensities can be used and are advisable to provide a safety margin in case the efficacy of nerve stimulation decreases during surgery. Stimulus rate Rapid stimulus delivery rates should be avoided, as they degrade the SSEP waveforms. In clinical settings, a rate of 3-6 stimuli per second usually is used. Rates that are exact subharmonics of 60 Hz (eg, 5 Hz) should be avoided, since their use would lead to contamination of the averaged SSEPs by large artifacts of the line frequency (50 or 60 Hz). RECORDING PARAMETERSSection 4 of 10
SSEPs typically are recorded by using standard EEG electrodes affixed with tape or collodion; electrode caps containing multiple recording electrodes also can be used. Scalp needle electrodes are not used routinely because of their high impedance, risk of infection, and discomfort to the patient. Recording electrode impedances should be kept below 5000 ohms and should be as uniform as possible across the electrodes to maximize common-mode rejection and minimize noise pickup. Also, placing the ground electrodes on the stimulated limb, proximal to the stimulation site, helps to reduce the electrical stimulant artifact. Typical recording amplifier filter settings for SSEPs are 30-3000 Hz. Diagnostic SSEP studies should be performed using the same filter settings that were used to record the normative data. Small-amplitude components of SSEPs are composed of both low and high frequencies, and filtering can be problematic. A bandpass that is too wide results in noisy SSEPs, but a bandpass that is too restrictive attenuates either high- or low-frequency components, depending on the settings chosen, and distorts the SSEP waveform. For example, reducing the low-frequency filter setting (low-cut, high-pass) from 30 to 5 Hz may produce a clearer SSEP component but also may allow more noise into the SSEP waveforms. A typical analysis time is 40 milliseconds for an upper limb SSEP and 60-80 milliseconds for a lower limb SSEP. Typically, SSEPs are not visible in the raw data recorded from surface electrodes, and signal averaging is used to extract the SSEPs from the other electrical signals picked up by the recording electrodes. Online artifact rejection is used to prevent epochs with unusually high noise levels from contaminating the averages. Sedation also may be used to reduce artifacts due to muscle and patient movement. Sleep causes minor changes in the SSEP waveforms. SSEPs do not have a uniform polarity convention. A negativity at input 1 with respect to input 2 is usually displayed as an upward deflection in the waveform. MEASUREMENT OF SOMATOSENSORY EVOKED POTENTIALSSection 5 of 10
Several characteristics of SSEPs can be measured, including peak latencies, interpeak latencies, morphology (ie, presence and absence of components), and dispersion. Peak latencies are the easiest SSEP feature to measure and standardize. Other characteristics (ie, morphology and dispersion) are more variable and difficult to interpret. Absolute SSEP latencies vary with limb length. Interpeak (ie, transit) times are useful because they separate out effects of limb length and of peripheral nerve disease, which may prolong absolute SSEP latencies in the absence of central nervous system pathology. Aging is associated with some prolongation of SSEP latencies. Latencies are considered abnormal when they are more than either 2.5 or 3 standard deviations above the mean of the normative data (depending on the laboratory's conventions). Recording electrodes sites Recording electrode sites are identified by anatomical landmarks. Those on the head are defined using the international 10-20 system, or its extension, the 10-10 system. Electrode CP3 is midway between C3 and P3, and electrode CP4 is midway between C4 and P4. CPi denotes either CP3 or CP4, whichever is ipsilateral to the stimulated limb; CPc is the contralateral centroparietal scalp electrode. CPz is midway between Cz and Pz. Recording electrodes over the spine are placed in the midline, and they are labeled with the name of the vertebral body they are placed on followed by the letter S, for example, T10S. Recording montages for cortical SSEP components are either cephalic bipolar, in which both electrodes are placed over the head, or referential, in which a reference electrode is placed at a noncephalic site. Cephalic bipolar montages have the advantage of being relatively free from noise and are preferred for routine clinical use. For upper limb SSEP studies, electrodes are placed over Erbs point (ie, the angle between the clavicular head of the sternocleidomastoid muscle and the clavicle), both ipsilateral and contralateral to the stimulus (labeled EPi and Epc). For lower limb SSEP studies, IC denotes an electrode placed over an iliac crest. Montages for the upper limb SSEPs The minimal montage recommended by the American EEG Society (AEEGS, currently the American Clinical Neurophysiology Society) for recording of upper limb SSEPs is as follows:
An example of a median nerve SSEP recorded with this montage is shown in Image 1. Montages for lower limb SSEPs The minimal montage recommended by the AEEGS for recording of lower limb SSEPs is as follows:
An example of a posterior tibial nerve SSEP recorded with this montage is shown in Image 2. This montage does not include a channel for recording the peripheral nerve CAPs at the popliteal fossa. Thus, optimal recording of lower limb SSEPs requires more than 4 channels; a signal averager with at least 8 input channels offers the advantage that all channels can be recorded simultaneously. Spinal SSEP Spinal SSEPs are recorded by electrodes placed over the spine as described above. They are considerably smaller in amplitude than SSEPs recorded over the scalp. However, the difference in latency between the scalp and the cervical or limb SSEPs is a measure of central sensory conduction, assessment of which remains the chief clinical goal of recording SSEPs. Thoracolumbar spinal SSEPs are even smaller than cervical spinal SSEPs and can be difficult to record, especially in obese subjects. COMPONENTS OF SOMATOSENSORY EVOKED POTENTIALSSection 6 of 10
SSEP components typically are named by their polarity and typical peak latency in the normal population. For example, N20 is a negativity that typically peaks at 20 milliseconds after the stimulus. The normal latency value for a component in a particular individual may be different from that implied by the component's name, because the lengths of the peripheral nerve and spinal conduction pathways, which vary with the patient's stature, influence the latencies of the SSEP components. The nomenclature of an evoked-potential component is inconsistent in the literature because the recording montage is not specified with the peak latency and polarity. For example, a CPi-EPc linkage following median nerve stimulation records a P14, while an EPc-CPi linkage records an N14. Upper limb somatosensory evoked potentialsPeripheral nerve compound action potential During clinical diagnostic studies of the upper limb SSEP, a surface electrode at Erbs point is used to record the peripheral nerve CAP as it traverses the brachial plexus. N9, the initial negative peak (labeled EP in Image 1), reflects the CAP within the most rapidly conducting subset of the afferent fibers. Multiple negative peaks, reflecting peripheral nerve fiber populations with different conduction velocities, may be recorded in some subjects, most often in children. When this occurs, the earliest negative peak should be interpreted as the N9 peak. A smaller P9 far-field peak, which most likely also arises within the brachial plexus, may be seen in scalp-to-noncephalic recordings; it has a slightly shorter latency than N9. Erbs point-recording electrodes have several disadvantages during intraoperative monitoring, such as their proximity to the sterile field, ease of dislodgment, and pickup of ECG artifact. A useful alternative recording site is over the peripheral nerve in the antecubital fossa (see Image 3). Cervical components An SSEP component that most likely arises in the first-order afferent neuron at or near the dorsal root entry zone (ie, in the dorsal root and/or the dorsal column) can be recorded as a far-field P11 peak in scalp-to-noncephalic reference recordings and as a near-field N11 peak in surface recordings over the lower cervical spine. This component is small and is not identifiable in all healthy subjects. A larger and more consistent component recorded over the lower cervical spine (eg, at SC5 or SC7) is N13. N13 has a horizontally oriented voltage field, negative dorsally and positive ventrally (see Image 4), and is generated by postsynaptic activity of neurons in the gray matter of the lower cervical spinal cord. It sometimes is called the stationary cervical potential, because its latency is not affected by cervical recording electrode location. Far-field components The stationary cervical potential overlaps in time with a far-field SSEP component, P14. The dipole orientation of P14 is such that it appears as a positive peak in recordings between the dorsal scalp (input 1) and a noncephalic electrode (input 2) (see Image 1). While the origin of P14 has been the subject of some controversy, it most likely reflects activity in the dorsal column nuclei and/or the caudal medial lemniscus within the lower medulla. When a forehead (ie, Fpz) reference is used, this far-field cervicomedullary component becomes a negativity (N14) at the C5S recording location and summates with the near-field N13 negativity picked up by that dorsal neck electrode. For intraoperative monitoring, the cervicomedullary far-field potential may be recorded at the inion, mastoid, or earlobe, referred to as Fpz. It appears as a negative peak, N14, in these recording linkages and, importantly, it is not contaminated by the N13 cervical near-field potential. N14 can be used to determine whether activity in afferent somatosensory pathways reaches the level of the cervicomedullary junction (see Image 5). Since at least 2 more synapses (in thalamus and cortex) intervene, the N14 component may permit SSEP monitoring of the cervical spinal cord when cortical SSEPs are of poor quality because of high anesthetic levels and/or preexisting neuronal damage. If the region of the nervous system in jeopardy is rostral to the cervicomedullary junction, N14 can be monitored to determine whether changes in the cortical SSEPs are due to rostral nervous system dysfunction, to peripheral nerve, or to technical problems. This is similar to the intraoperative use of the peripheral nerve SSEP component described above. Optimally, both components should be monitored for 2 reasons: (1) N14 provides an alternative way of differentiating the possible causes of a cortical SSEP change if peripheral nerve SSEP recordings are suboptimal, and (2) if peripheral nerve CAPs are interpretable and remain unchanged while cortical SSEPs deteriorate, examination of the N14 recordings can localize further the neural dysfunction responsible for the cortical SSEP changes above or below the cervicomedullary junction. Another far-field component, N18, overlaps in time with the primary cortical SSEP and may account for multiple negative peaks in the cortical recordings in some subjects (see Image 1). N18 has a wide bilateral distribution over the scalp. It is best seen in recordings with a noncephalic reference, though it also may be demonstrated with a frontal reference. While N18 has been attributed to a thalamic generator, several cases have been reported in which N18 was still present despite the presence of thalamic lesions that eradicated the primary cortical SSEP. N18 most likely reflects activity in multiple subcortical structures that are activated by the somatosensory stimulus, including brainstem structures. Thus, examination of N18 cannot be used to localize the cause for cortical SSEP changes as rostral versus caudal to the thalamus. Cortical components The primary cortical SSEP component following median nerve stimulation, N20 (see Image 1), is recorded as a near-field potential over the parietal area contralateral to the stimulated median nerve. Since an electrode also is within the scalp distribution of the far-field N18 component, a recording with a noncephalic reference contains an admixture of N18 and N20. The recommended montage places the reference over the mirror-image position ipsilateral to the stimulus (ie, a CPc-CPi recording linkage), which tends to cancel the bilaterally distributed N18. An Fz reference also has been used to record N20 but may yield a composite waveform consisting of N20 and the inversion of the frontally generated P22 component. A CPc-Fz or CPc-Fpz linkage may be useful for intraoperative monitoring because it yields a cortical SSEP waveform that is larger in amplitude than the CPc-CPi linkage used for diagnostic recordings. While a thalamic or subcortical origin for N20 has been suggested, most authors believe that N20 predominantly reflects activity of neurons in the hand area of the primary somatosensory cortex; multiple generators with overlapping voltage topographies may contribute to this component. N20 predominantly originates in primary somatosensory cortex in the posterior bank of the central sulcus (see Image 6) and thus displays a polarity inversion across the central sulcus in epidural cortical surface recordings (see Image 7) and some scalp recordings. This polarity inversion may be used to identify the central sulcus during surgery. An accumulation of air within the subarachnoid space can block the volume conduction of the cortically generated SSEP components to the overlying recording electrodes; this is most likely when the patient is in the sitting position and cerebrospinal fluid (CSF) has been drained. The peripheral nerve and far-field cervicomedullary SSEPs are not affected generally. Therefore, the SSEP changes produced by pneumocephalus can mimic those caused by intraoperative compromise of the somatosensory pathways rostral to the cervicomedullary junction. Lower limb somatosensory evoked potentialsPeripheral nerve compound action potential A surface electrode placed in the popliteal fossa in the midline can be used to record the peripheral nerve CAP following posterior tibial nerve stimulation at the ankle. To minimize both electrical stimulus artifact and ECG pickup, a reference electrode on the same leg is used. Possible linkages include a midline electrode 2 cm above the popliteal crease referred to a midline electrode 5 cm above the popliteal crease and a midline popliteal fossa electrode referred to an electrode placed at the lateral aspect of the same knee. In patients in whom a clear foot twitch is not obtained, the presence of a clear peripheral nerve CAP at the popliteal fossa demonstrates that the posterior tibial nerve has been stimulated adequately. In this case, the absence of more rostrally generated SSEP components is evidence of abnormality within the neural somatosensory pathways. Without the peripheral nerve recording channel, the absence of SSEPs also could have been due to technical factors that prevented adequate nerve stimulation. Similarly, during intraoperative monitoring, examination of the peripheral nerve recording channel permits a rapid classification of the causes of changes in the more rostrally generated SSEP components, in the same manner as was described above for upper limb SSEPs. Lumbar components An electrode placed over the lower thoracic or upper lumbar spine records a combination of the CAP in the primary afferent neuron, propagating within the cauda equina and fasciculus gracilis, and a stationary lumbar potential (SLP) that is derived from postsynaptic neurons in the gray matter of the spinal cord. Recordings with a distant reference, such as the iliac crest, emphasize the SLP, which is analogous to the stationary cervical potential (N13) recorded over the lower cervical spine following median nerve stimulation. Bipolar recordings between a pair of rostrocaudally separated electrodes over the lower spine record the propagating CAP. However, they also contain a component derived from the SLP, representing the difference in the amplitude of the SLP between the 2 recording electrodes. The relative magnitudes of the CAP and SLP contributions vary across subjects; therefore, referential recordings show less intersubject latency variability and should be used for clinical diagnostic SSEP testing. The normal amplitude maximum of the SLP is at the T10-T12 vertebral level but may be fairly restricted (see Image 8). Thus, referential recordings from multiple electrodes over the lumbar and lower thoracic spine may be useful in demonstrating the SLP. The following channels are included in an extended montage suggested in the AEEGS guidelines:
In patients with tethered spinal cords, recordings from such an array of electrodes often demonstrate caudal displacement of the maximal amplitude of the SLP, reflecting the anatomical displacement of the lower spinal cord, or may show no identifiable SLP. The lumbar SSEP components are sometimes not identifiable in unsedated healthy subjects, especially if they are obese. Far-field components When referred to a frontal scalp reference, electrodes over the cervical spine record a biphasic waveform that was labeled originally as a cervical potential but is believed now to reflect predominantly far-field potentials generated in subcortical elements of the lemniscal somatosensory pathways. If the inputs are connected so that the cervical lead is input 1 (ie, C5S-Fpz), the waveform consists of a negativity followed by a positivity. The montage given in the AEEGS guidelines specifies this channel as an Fpz-C5S recording, with Fpz as input 1. Such a recording contains a positivity followed by a negativity, which are designated P31 and N34, respectively, following posterior tibial nerve stimulation (see Image 2); the peak latencies are approximately 10 milliseconds shorter, following peroneal nerve stimulation. P31 is analogous to the P14 component of the median nerve SSEP and, like it, most likely is generated in the dorsal column nuclei and/or the caudal medial lemniscus within the lower medulla. The negativity that follows is most likely analogous to the N18 component of the median nerve SSEP and thus is generated by activity in multiple brainstem and/or thalamic structures that are activated by the somatosensory stimulus. A near-field origin within the cervical spinal cord had been proposed for some smaller and earlier peaks that can be picked up by the recording electrodes placed on the posterior part of the neck. Like the lumbar SSEP components, the far-field SSEP components elicited by lower limb nerve stimulation may be difficult to identify in recordings from unsedated subjects because of noise, particularly electromyography (EMG) artifact from paraspinal musculature. Under surgical anesthesia, and especially with the use of neuromuscular blocking agents, they are usually clearly identifiable and reproducible. The cortically generated SSEPs due to stimulation of lower limb nerves are far more sensitive to the effects of anesthesia than the far-field components (see Image 9). During operations in which the cortically generated SSEPs are markedly attenuated or completely suppressed by anesthesia or in which they show a degree of anesthetic-related variability such that changes related to surgical manipulations might not be recognized, the far-field SSEPs may be used to monitor the dorsal column pathways of the spinal cord. Cortical components The primary cortical SSEPs following lower limb stimulation are recorded as near-field positivities: P37 following posterior tibial nerve stimulation and P27 following peroneal nerve stimulation. In contrast to the N20 cortical component of the median nerve SSEP, which is maximal over the lateral parietal area, the cortical SSEPs to lower limb nerve stimulation often are maximal near the midline, reflecting the more medial location of the foot and leg areas of the somatosensory homunculus. However, the topography of these components varies across subjects, reflecting individual differences in the location and orientation of somatosensory cortex (see Image 10) and may differ between the 2 sides in the same subject. When the equivalent dipole of the cortical generator is oriented vertically, the maximum positivity is in the midline (see Image 10A, Image 11A). When the cortical generator is located in the mesial wall of the hemisphere, the dipole assumes a more horizontal orientation, producing a paradoxical maximum positive cortical SSEP over the hemisphere ipsilateral to the stimulus (see Image 10B, Image 11B).
When the dipole is oriented horizontally, producing an ipsilateral positivity and a contralateral negativity, a CPi-CPc recording channel may yield a larger cortical SSEP than any of the recordings referred to Fpz. Thus, some laboratories incorporate a Cpi-CPc derivation in their recording montage. However, the latencies of the ipsilateral positivity and the contralateral negativity may not be identical, and the Cpi-CPc SSEP may be a composite potential reflecting multiple cortical sources. Thus, a Cpi-CPc SSEP should be compared only to normative data recorded with the same derivation. Since the largest cortical SSEP is located at CPz in some patients, while in others with a more horizontal dipole a small cortical SSEP is recordable at this electrode but a substantial SSEP can be recorded over the ipsilateral hemisphere, the cortical SSEP recording electrode sites used for intraoperative monitoring should be customized on the basis of the results of each patient's preoperative SSEP studies. Therefore, these studies must include recordings of the cortical SSEPs from both midline and lateral electrodes. With both peroneal nerve and posterior tibial nerve stimulation, longer latency cortical SSEP components follow the primary cortical P27 or P37 components. The second cortical positivity, which typically has a latency of 50-60 milliseconds after posterior tibial nerve stimulation, may be substantially larger than the primary cortical SSEP (see Image 12). In noisy recordings with a limited montage, a low-amplitude P37 component may not be recognized, and the secondary cortical positivity may be identified erroneously as a markedly delayed cortical SSEP (see Pz-Fpz waveforms in Image 12). Thus, cortical SSEPs to lower limb nerve stimulation should be interpreted with caution when their peak latencies appear to be delayed abnormally to this latency range. NEAR-FIELD SPINAL SOMATOSENSORY EVOKED POTENTIALSSection 7 of 10
In some patients who are undergoing surgery for lesions that have already caused spinal cord dysfunction, both cortical and far-field SSEPs may be absent in surface recordings because of a combination of anesthetic effects and temporal dispersion of the afferent volley. Electrodes positioned close to the spinal cord may pick up reproducible near-field SSEPs, permitting intraoperative monitoring. Because the spinal cord volley in these patients is desynchronized and/or because the near-field electrode picks up activity in several different fiber populations, the SSEPs picked up by electrodes on the dorsal surface of the spinal cord often are complex and polyphasic. This may preclude simple latency and amplitude measurements, but SSEP changes caused by manipulation of the spinal cord or irrigation with cold fluids still can be recognized (see Image 13). INTERPRETATION OF SOMATOSENSORY EVOKED POTENTIALSSection 8 of 10
Spinal cord pathways mediating the somatosensory evoked potentials The large-fiber, rapidly conducting afferent somatosensory pathways that sustain the primary cortical SSEPs to stimulation of mixed sensorimotor limb nerves travel predominantly in the dorsal columns within the spinal cord. In experimental animals, transection of the dorsal column pathways almost completely obliterates the earliest cortical SSEPs to stimulation of more caudally located peripheral nerves, while ventrolateral funiculus lesions usually have only minor effects on these SSEPs. Thus, significant damage to descending motor systems can occur without causing changes in the SSEPs used for intraoperative monitoring. Such false-negative cases are fortunately rare, but they have occurred. In contrast to the cortical SSEPs, near-field SSEPs recorded over the spinal cord (see Image 13) may contain components reflecting large-fiber afferent activity in both the dorsal columns and the spinocerebellar tracts. Clinical diagnostic studies SSEP amplitudes may vary considerably across subjects, and the interpretation of clinical diagnostic SSEP studies is based primarily on component latencies. For upper limb SSEPs, the AEEGS guidelines include identification of the obligate components N9, N13, P14, N18, and N20. Measurement of the N9-N20, N9-P14, and P14-N20 interpeak intervals is specified; interpeak intervals involving the N13 component are listed as options. The N9-P14 interpeak interval measures neural conduction from the brachial plexus to the lower brain stem, P14-N20 from the lower brain stem to the primary somatosensory cortex, and N9-N20 from the brachial plexus to the primary somatosensory cortex. The N13 component reflects activity within the lower cervical spinal cord. Absolute component latencies are less useful than interpeak intervals because of the effect of arm length on N9 latencies. For posterior tibial nerve SSEPs, the AEEGS guidelines mandate at minimum the identification of the SLP and the primary cortical SSEP component (P37), measurement of their peak latencies, and calculation of the SLP-P37 interpeak interval. The latter approximates the conduction time between the lumbar spinal cord and primary somatosensory cortex. Some laboratories also measure the latency of the initial far-field component (P31) and calculate SLP-P31 and P31-P37 interpeak latencies, corresponding to lumbar spinal cord-to-brain stem and brain stem-to-primary somatosensory cortex conduction times, respectively. For lower limb SSEPs, conduction distances along both the peripheral nerve and the spinal cord portions of the afferent pathway vary considerably among subjects as a function of height. Some laboratories use height-adjusted norms for the analysis of lumbar SSEP components; others measure the conduction distances and calculate conduction velocities. Correction for the patient's height is more important for evaluating the absolute latency of P37 than the SLP-37 interpeak latency. When no height corrections are used, interpeak latencies should be interpreted with caution in patients whose heights are at the extremes of the range of heights for which the normative data were collected. The major criteria for abnormality are absence of obligate components and abnormal prolongation of interpeak intervals. As with any evoked-potential test involving unilateral stimulation, excessive symmetries between the measurements following left-sided stimulation and those following right-sided stimulation also may reveal abnormalities. The stationary cervical/lumbar and far-field components may be difficult to identify in some healthy subjects who are not sedated. In the presence of normal cortical SSEPs, the inability to identify some of the more caudally generated SSEP components may not be significant. SSEPs in comatose patients In adults who are comatose following a cerebral anoxic insult (eg, due to a cardiopulmonary arrest), bilateral absence of the N20 (cortical) component of the median nerve SSEP predicts a bad outcome (death or persistent vegetative state) with virtually 100% specificity. Preservation of the SSEPs does not, however, ensure a good outcome, as patients may have extensive brain injury despite preservation of the afferent somatosensory pathways and primary somatosensory cortex, or they may die from other disease processes (eg, cardiac disease that caused the arrest). Bilateral absence of the N20 is also associated with poor outcomes in adult patients with coma due to traumatic brain injury, although some of these patients may recover consciousness if the absence of SSEPs was due to a focal lesion that interrupted the pathway between the peripheral nerve and somatosensory cortex. Patients comatose from other causes such as infection and metabolic derangements may also have better outcomes despite bilateral absence of N20 on initial examination. Intraoperative monitoring A comprehensive discussion of the interpretation of intraoperative evoked potential data is beyond the scope of this presentation and the reader is referred to other sources. When surgical maneuvers compromise neural tissue, SSEP components may show significant amplitude attenuation before their latencies become prolonged. Thus, both amplitudes and latencies should be evaluated during intraoperative monitoring. No universally accepted standard exists for what constitutes a significant change, but a 50% decrease in the amplitude of an SSEP component or a 10% increase in its latency often are used as threshold criteria (SSEP amplitudes generally show more run-to-run variability than SSEP latencies). Whenever the SSEPs change, the interpreter must distinguish between the many possible causes of such a change, which include anesthetic effects and technical factors as well as true neuronal damage or dysfunction. As discussed above, recordings of peripheral nerve CAPs and cervicomedullary far-field potentials can help elucidate the causes and significance of changes in cortical SSEPs during intraoperative SSEP monitoring for both upper limb and lower limb SSEPs. Anesthetic agents are probably the most common cause of intraoperative SSEP changes. In general, the longer the latency of an SSEP component and the more synapses between the stimulation site and the component's neural generator, the greater is the degree to which that component will be affected by anesthetic agents. Thus, anesthetic effects may alter the cortical SSEPs while sparing the far-field SSEPs (see Image 5, Image 9), mimicking surgery-related dysfunction of the cerebral cortex or of the pathways from the brain stem to the cerebral cortex. Personnel performing intraoperative monitoring must pay careful attention to the anesthetic regimen and should record it periodically on their data logs. Personnel performing intraoperative monitoring also should periodically note and log the temperature and blood pressure of the patient, which also can affect the electrophysiologic signals. Anesthetic-induced changes typically are bilateral; this can help distinguish anesthetic-related from surgery-related SSEP changes when the latter are expected to be unilateral but not when surgical manipulations can damage afferent sensory pathways bilaterally (eg, bilateral spinal cord damage during surgery for scoliosis). SSEPs in children The cortical SSEP to posterior tibial nerve stimulation may be absent in healthy infants as old as 3 months. The cortical SSEP to median nerve stimulation also may be absent at birth but most likely is present consistently in healthy infants at an earlier age than the corresponding component of the lower limb SSEP. SSEP component latencies are, in general, shorter in infants and children than in adults and change progressively with growth and maturation. The latency changes predominantly reflect linear growth with elongation of the peripheral nerves and central somatosensory pathways. These effects are counterbalanced partially by myelination and increase in the fiber diameters, which produce faster conduction velocities, and partially by maturation of synaptic transmission. The latter effects operate until age 6-8 years, at which time central conduction times have reached adult levels and further latency changes are due to changes in stature. During spinal surgery in children in whom cortical SSEPs are absent or easily attenuated by anesthesia, the far-field SSEP components may be used to monitor the dorsal column pathway of the spinal cord. However, many patients with lumbar meningomyeloceles have conduction abnormalities (eg, conduction blocks, temporal dispersion) at the level of the neural plaque so that both far-field and cortical SSEPs are absent; this precludes intraoperative assessment of their dorsal column pathways. MULTIMEDIASection 9 of 10
REFERENCESSection 10 of 10
Somatosensory Evoked Potentials: General Principles excerpt Article Last Updated: May 15, 2006 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
