ACNS Guideline

Motor evoked potentials (MEPs) are electrical signals recorded from neural tissue or muscle after activation of central motor pathways. They complement other clinical neurophysiology techniques, such as somatosensory evoked potentials (SEPs), in the assessment of the nervous system, especially during intraoperative neurophysiologic monitoring (IONM). Somatosensory evoked potentials directly assess only a part of the spinal cord, the dorsal columns (Emerson, 1988), and also the medial lemniscus, the thalamocortical radiations, and somatosensory cortex. Because they provide indirect surveillance of the motor tracts, their use has been shown to improve neurologic outcomes during spinal surgery (Nuwer et al., 1995). However, SEPs can fail to detect damage to the spinal cord motor pathways when the dorsal columns are spared (Ben-David et al., 1987; Ginsburg et al., 1985; Jones et al., 2003; Krieger et al., 1992; Legatt et al., 2014; Zornow et al., 1990); this led to the development of techniques for directly monitoring the central motor pathways. Most often, this is accomplished using transcranial electrical stimulation (TES) of the brain and recording of evoked neural or myogenic activity caudal to the area that is at risk during surgery (Legatt, 2002). During TES, high-intensity stimuli must be delivered to the scalp to stimulate the brain through the intact skull, with stimulus voltage and current levels far above those used to elicit SEPs. If a craniotomy permits direct stimulation of motor cortex by electrodes placed on the brain surface, low-intensity direct cortical stimulation can also be used to elicit MEPs for IONM (Szelényi et al., 2007b; Taniguchi et al., 1993). Direct cortical stimulation is outside the scope of this guideline, but the recommendations herein for the recording of the MEPs that are elicited by transcranial electrical brain stimulation would also apply to recording of MEPs elicited by direct cortical stimulation. Transcranial magnetic stimulation has also been used to elicit MEPs by inducing electrical current flows within the brain tissue without passing large amounts of current through the scalp. This reduces stimulation of pain fibers in the scalp, skull, and meninges and makes it a practical technique for MEP studies in awake subjects (Chen et al., 2008). However, transcranial magnetic stimulation is not the optimal MEP technique for IONM because of the anesthetic suppression of transcranial magnetic stimulation–MEPs which are generated mainly by eliciting I-waves (see section on Definitions and Physiology, below) and difficulties in maintaining a constant position of the coil relative to the patient's head (Legatt, 2004). Neither TES with single stimulus pulses nor transcranial magnetic stimulation consistently produces robust myogenic MEPs suitable for IONM. The commercial availability of stimulators that can deliver trains of high-intensity electrical pulses has made reliable MEP monitoring using TES possible in most patients. At this time, the techniques for recording and interpreting TES-MEPs have become sufficiently well established to warrant the formulation of these guidelines. Personnel performing TES-MEP monitoring must be cognizant of the technical challenges and risks of the technique. TERMINOLOGY: DEFINITIONS AND PHYSIOLOGY Corticospinal tract activity elicited by stimulation of cerebral cortex, either electrical or magnetic, consists of “D-waves,” which reflect direct activation of the pyramidal cell axons that leave the cortex and comprise the corticospinal tract, and “I-waves,” which reflect indirect activation of these pyramidal neurons by synaptic transmission from activated cortical interneurons (Amassian et al., 1987) (Fig. 1). There may be multiple I-waves, at roughly equal intervals, reflecting the number of synapses (and synaptic delays) between the interneurons that are initially activated by the stimulus and the pyramidal neurons that give rise to the corticospinal tract. Since I-waves are mediated by cortical synaptic activity, they are markedly suppressed by surgical levels of anesthesia. However, D-waves remain and can be recorded along the course of the corticospinal tract. When used for IONM, they are recorded from the spinal cord caudal to the region that is at risk during the operation. Recordings of D-waves from the spinal cord rostral to the region at risk can also be performed as a control to assess the adequacy of corticospinal tract stimulation.FIG. 1: Corticospinal tract activity elicited by stimulation within or below motor cortex, recorded from the ipsilateral lateral column between C1 and C2, in a monkey. Intracortical stimulation (upper trace) produces a D-wave and a series of I-waves; stimulation within the subcortical white matter (lower trace) only produces a D-wave (Modified from Patton HD, Amassian VE. Single- and multiple-unit analysis of cortical stage of pyramidal tract activation. J Neurophysiol 1954;17:345–363).When brain stimulation causes muscle contractions, the compound muscle action potentials, or myogenic MEPs, can be recorded from multiple muscles simultaneously following a single train of transcranial stimuli. Train stimulation is needed to reliably elicit myogenic MEPs under anesthesia. The excitatory postsynaptic potentials in the anterior horn cells summate to bring them to threshold and fire them. Both D-waves and myogenic MEPs can be used for IONM (Fig. 2).FIG. 2: Concurrent monitoring of myogenic motor evoked potentials (MEPs) (left) and D-waves (right) during surgery for a spinal cord tumor. The myogenic MEPs disappeared, but the D-waves persisted. The patient awoke with new neurologic deficits that subsequently cleared (From Deletis V, Kothbauer K. Intraoperative neurophysiology of the corticospinal tract. In: Stålberg E, et al., eds. Spinal cord monitoring. Wien: Springer, 1998:421–444).Transcranial electrical stimulation predominantly generates D-waves under the stimulating anode and therefore predominantly generates myogenic MEPs in muscles contralateral to the stimulating anode (Fig. 3). However, in rare patients with congenital motor tract nondecussation, TES produces predominantly anode-ipsilateral myogenic MEPs (MacDonald et al., 2004). Myogenic MEPs may also be elicited by stimulation of cortex under the TES cathode, but such responses are less stable and may disappear in the absence of corticospinal tract pathology (Legatt, 2006).FIG. 3: Motor evoked potentials to transcranial electrical stimulation between electrodes at positions C1 and C2, recorded bilaterally from thenar and tibialis anterior muscles during an occipitocervical fusion.Mention should be made of a technique using stimulation of the rostral spinal cord and recordings from peripheral nerves in the legs, which has sometimes been labeled “neurogenic motor evoked potentials” (Owen et al., 1988). Collision studies (Toleikis et al., 2000) have shown that the signals recorded using this technique are mediated by retrograde conduction within the dorsal columns, not anterograde conduction within the corticospinal tracts. Also, these signals may be preserved in the face of corticospinal tract damage that causes paraplegia (Minahan et al., 2001). This technique may be useful for IONM of the dorsal columns, but it should not be construed to be an MEP monitoring technique. Similarly, myogenic MEPs after rostral spinal cord stimulation could be partly mediated through retrograde activation of the dorsal columns, whose collateral branches form excitatory synapses with alpha motor neurons (MacDonald, 2006), and recording of these signals should not be considered a reliable method for corticospinal tract monitoring. Comparison of D-wave and Myogenic MEP Monitoring D-waves to single-pulse TES are typically recorded as they pass through the corticospinal tract within the spinal cord using near-field electrodes, such as epidural or subdural electrodes. Because there are no synapses between the stimulated cortical pyramidal neurons and the MEP recording site, multipulse stimulation is not required, although a high stimulus intensity is still required to stimulate the brain through the intact skull. The lack of synapses makes D-waves relatively insensitive to anesthesia. D-waves tend to be highly consistent from run to run but are generally small enough to require averaging a small number of responses (≤20) per run to improve the signal-to-noise ratio. The D-wave amplitude corresponds to the number of rapidly conducting corticospinal tract axons within the spinal cord at the level of the recording. Since some corticospinal tract axons terminate at each segmental level, D-waves are of higher amplitude in the cervical region than in the thoracic spinal cord. D-wave monitoring is usually not practical below the T10 bony level because of the small number of corticospinal tract fibers that remain. Myogenic MEPs are large and do not require signal averaging. Moreover, they often display substantial run-to-run variability (Fig. 4), so that signal averaging should not be used to record them. Generating a myogenic MEP requires synaptic transmission at the anterior horn cell, which is facilitated by a train of stimulus pulses. A train of multiple stimuli also facilitates the production of I-waves, further increasing the potency of the train (Deletis et al., 2001b). The interposed synapse at the anterior horn cell makes myogenic MEPs highly sensitive to anesthetic effects (Fig. 5) and likely accounts for most of their run-to-run variability (MacDonald, 2006). Each stimulus train activates only a small fraction of the anterior horn cells; a varying subset of the lower motor neuron pool is recruited with each run, causing the variability in the response waveforms from run to run.FIG. 4: Myogenic motor evoked potentials recorded from the left tibialis anterior and thenar muscles after multipulse transcranial electrical stimulation with the anode over the right hemisphere, over a 3-hour period during an occipitocervical fusion. Note the large run-to-run variability of the motor evoked potential amplitudes and waveshapes. The numbers in the middle are the clock times of each run (From Legatt AD, Ellen R. Grass Lecture: Motor evoked potential monitoring. Am J Electroneurodiagnostic Technol 2004;44:223–243).FIG. 5: Effect on different evoked potentials of changing the anesthetic regimen from total intravenous anesthesia with propofol and remifentanil to inhaled sevoflurane during spinal instrumentation surgery for a T11 compression fracture with kyphosis. Subcortical somatosensory evoked potentials (SEPs) (left column, arrow) were essentially unaffected, cortical SEPs (middle column) were attenuated, and myogenic motor evoked potentials (MEPs) (right column) were almost completely eliminated. TCES = trans-cranial electrical stimulation (From Legatt AD. Intraoperative neurophysiology: Interactive case studies. New York: Demos Medical Publishers, 2015).D-waves, myogenic MEPs, or a combination of both may be used for monitoring the spinal cord. Each has advantages and disadvantages (Legatt, 2004), as described in the following list. In the United States, most centers routinely perform myogenic MEP monitoring. Anesthesia: D-waves are relatively insensitive to anesthesia. Myogenic MEPs are easily suppressed by anesthesia, especially by inhalational anesthetics (Fig. 5), which sets limits on the anesthetic regimen that can be used during MEP monitoring of myogenic MEPs (Sloan and Heyer, 2002). Neuromuscular blockade (NMB): neuromuscular blockade does not affect D-waves, but total NMB eliminates myogenic MEPs. Omitting NMB permits straightforward myogenic MEP monitoring but TES-induced patient movements may necessitate careful stimulus timing to avoid unacceptable patient movements that can interfere with the surgery. Partial NMB may dampen but not eliminate these movements and can complicate interpretation of myogenic MEPs. Some centers generally omit NMB whereas others tend to use partial NMB. If used, partial NMB should be done with continuous infusion of the paralytic drug (Adams et al., 1993); bolus injections yield a level of NMB that is too variable. Stimulator: monitoring of myogenic MEPs requires a multipulse stimulator; D-wave monitoring does not. Recording electrodes: D-wave monitoring requires invasive electrodes placed near the spinal cord, either intraoperatively by the surgeon or percutaneously; monitoring of myogenic MEPs does not. Structures monitored: since D-waves must be recorded caudal to the region at risk for IONM, they cannot be used to monitor the lower spinal cord (usually caudal to the T10 bony level). D-waves assess only axonal conduction within the corticospinal tracts. Myogenic MEPs additionally assess the integrity of spinal cord gray matter, which may be more sensitive to ischemia than spinal cord white matter (MacDonald and Dong, 2008) and may also demonstrate nerve root or peripheral nerve dysfunction. Desynchronized activity: spinal cord lesions such as tumors may cause temporal dispersion of the descending corticospinal tract volley, which precludes monitoring of D-waves caudal to the lesion. Myogenic MEPs adequate for monitoring may be present in such patients (Fig. 6). Detection of unilateral compromise: D-wave monitoring can fail to detect unilateral corticospinal tract compromise because the epidural electrodes record the D-waves generated in the corticospinal tracts on both sides. Myogenic MEP monitoring records from muscles on each side separately. Timing of alarm: because myogenic MEPs require not only corticospinal tract conduction but also anterior horn cell transmission and peripheral nerve conduction, they may be lost at a time when D-wave corticospinal tract potentials are still present (Fig. 2). The meaning of this dissociation depends on the surgical circumstances. During descending aortic surgery, acute spinal cord ischemia rapidly disables anterior horn cells, causing loss of myogenic MEPs, while corticospinal tract conduction and D-waves may be unaffected or begin to fail after a delay (MacDonald and Dong, 2008). In these surgeries, persistent loss of myogenic MEPs correlates with a substantial risk of cord infarction and permanent motor deficits (Keyhani et al., 2009). In contrast, during surgery for intramedullary spinal cord tumors, patients in whom myogenic MEPs are lost but D-waves persist generally have transient postoperative weakness or paralysis; the myogenic MEP changes may reflect disruption of propriospinal systems that render intact alpha motor neurons unexcitable, with functional compensation for the loss of these facilitatory inputs during the postoperative period (Deletis, 2002; Sala et al., 2006). Patients in whom D-waves are lost or attenuated by more than 50% during intramedullary spinal cord tumor surgery generally suffer permanent weakness (Deletis and Kothbauer, 1998). Thus, combined D-wave and myogenic MEP monitoring particularly suits these operations. Effect of spinal deformity correction: during correction of spinal deformities, changes in the anatomic relationship between the D-wave recording electrodes and the spinal cord can produce false-positive results during D-wave monitoring (Ulkatan et al., 2006). FIG. 6: No clear D-wave is present in epidural recordings caudal to an intramedullary tumor of the cervical spinal cord (left), but transcranial electrical stimulation with brief stimulus trains elicits clear myogenic motor evoked potentials (right) (From Deletis V, Kothbauer K. Intraoperative neurophysiology of the corticospinal tract. In: Stålberg E, et al., eds. Spinal cord monitoring. Wien: Springer, 1998:421–444).RECOMMENDED STANDARDS FOR TES-MEP MONITORING Standards for recording equipment, personnel, and documentation are given in the American Clinical Neurophysiology Society Guideline 9A: Guidelines on Evoked Potentials (2006). Additional standards for IONM are given in American Clinical Neurophysiologic Society Guideline 11: Guidelines for Intraoperative Monitoring of Sensory Evoked Potentials (1994a). Personnel involved in the interpretation of intraoperative MEP monitoring data should have additional training and experience that provides thorough understanding and direct familiarity with all aspects of TES and of MEP data acquisition, processing, and interpretation. This should include the influence of stimulus and recording parameters, anesthesia, neuromuscular blocking agents, and other factors that may affect the MEPs during IONM; knowledge of anatomic structures, neurophysiologic events and other factors involved in the generation of MEPs; the clinical significance and pathophysiologic correlates of dysfunction of neural pathways demonstrated by evoked potential alterations; and knowledge of which areas of the nervous system are at risk and the mechanisms for that risk during the surgical procedures for which MEP monitoring is used. Stimulating Equipment for TES Equipment for TES should be able to deliver brief trains of high-intensity stimuli where the intensity of the stimulus pulses, the number of pulses per train, and the interpulse interval (or equivalently, the pulse rate) within the train can all be adjusted by the operator. Either constant-voltage or constant-current stimulators can be used; specially designed devices of either type are available and some standard SEP stimulators of either type can also be effective (the latter is an off-label use of the SEP stimulator). In constant-voltage stimulation, the stimulus current can vary widely depending on the impedance of the tissue and the electrode–tissue interface. A display of the delivered current is desirable, and the equipment should include circuitry to limit the total current delivered during a stimulus train to a safe level. Isolation and leakage current limitations for evoked potential recording equipment (American Clinical Neurophysiology Society, 2006) also apply to TES stimulating equipment. If the TES is not within the MEP recording equipment, it must have a that can be to the MEP recording equipment to of stimulus with recording of the MEP Stimulating and or electrodes are most often used for TES, although electrodes can also be used. which with a area between and to tissue from the high stimulus used by current and to the tissue near the electrodes have lower than or electrodes (MacDonald, 2006) and are also less likely to become When MEPs in the are stimulating electrodes may be placed at scalp positions or of the system (American Society, with to stimulation of either the left or the right using the (Szelényi et al., (Fig. Since electrodes are to motor cortex, and nerves than are electrodes, stimulation at can produce and electrodes are electrodes are required to elicit MEPs (MacDonald, et al., Stimulating positions for transcranial electrical stimulation. The other positions within the different stimulating can be used to elicit MEPs in the lower electrodes at or can be used with MEPs are with these responses still tend to be contralateral to the anode (or ipsilateral in patients with anode at can be with a at (Fig. In some the anode is an at and the is a with a large area placed over the of the The and stimulating have the that they can stimulate the motor pathways for both and lower with a single stimulus train, recording of and MEPs. The stimulating has the that it may more reliably stimulate the motor pathways for the lower bilaterally with a single The optimal for stimulation may vary between patients and surgical different stimulation can be and the for each When myogenic MEPs are multipulse TES is used because under surgical levels of anesthesia, a single D-wave is often not to bring the anterior horn cell to the stimulation elicits a train of D-waves, and often some I-waves as and the excitatory postsynaptic potentials that they produce in the anterior horn cell summate to above the lower motor neuron and the myogenic MEP (Legatt, 2004). If the interpulse interval between stimulus pulses in the is too the postsynaptic potentials do not and the of the temporal is If it is too stimuli after the in the train are not as effective in the corticospinal tract axons because of their (Deletis et al., between and pulse of are typically optimal for myogenic MEP monitoring. A train of pulses in some others require Both constant-current and constant-voltage stimulators can be used for different have different of for pulse and stimulus Intraoperative neurophysiologic monitoring typically a standard of stimulus stimulus number of pulses per train, and interpulse rate) for the recordings under anesthesia, and the as and to MEPs adequate for IONM in each Transcranial electrical stimulation with of pulse trains et al., can the recording of myogenic MEPs, as can electrical stimulation of the recording of MEPs et al., these techniques are relatively for them are not in this Recording and Recording D-waves are recorded between electrodes placed near the spinal cord, either epidural or They may be placed using a or placed by the within the surgical there is a spinal cord such as a recording D-waves both rostral and caudal to the may be between the recording electrodes and produce D-wave amplitudes but more a of to is adequate (Deletis and 2008). Myogenic MEPs should be recorded from muscles on both of the and electrodes are both effective in recording these Either type and their should be to the to during surgery. In the myogenic MEPs are recorded from muscles or dorsal because of their predominantly under corticospinal tract especially can be used as MEPs are useful for monitoring the when the region at risk is above the lower cervical spinal cord, that the rostral cervical spinal cord, the or the corticospinal tracts within the When the thoracic and spinal cord are at MEPs may be used as a control recording to such as anesthesia, that be the MEPs recorded from lower They may also be used to monitor for compromise because of of the patient's In the lower the tibialis anterior and are the muscles most used for myogenic MEP monitoring. muscles may be used as but tend to give less reliable MEPs. Myogenic MEP recording should include muscles when the thoracic spinal cord is at Myogenic MEPs can also be recorded from the this is most often used during surgery on the lower spinal cord and in the region of the and on are in not useful during MEP monitoring (Deletis and 2008). are used to assess both D-waves and myogenic MEPs. The amplitude of the D-wave is from to the following of the is the case for IONM of evoked potentials, the most is a 50% in the signal The amplitude of the myogenic MEP is between the most and the most of the response to the variability of myogenic MEPs in the absence of spinal cord compromise (Fig. 4), a 50% amplitude is usually not an for spinal cord monitoring with myogenic MEPs, as it would cause too there is no as to an for myogenic MEP monitoring of the spinal the are still that is widely used during spinal cord tumor surgery is of the myogenic MEP in the threshold However, it may be to the of the surgical the myogenic MEP by a threshold than 50% or than it amplitude can but an still produces a number of during surgery et al., have been used during myogenic MEP an in the threshold stimulus intensity required to elicit an MEP et al., and a in the and of the myogenic MEP et al., et for a of MEP D-waves are relatively unaffected by anesthesia. However, to anesthetic effects at the synapse between the corticospinal tract and the anterior horn cell, myogenic MEPs are markedly by anesthesia, to a than most other used for IONM (Fig. the of the anesthetic regimen is particularly when myogenic MEPs are inhalational are because they myogenic MEPs, especially at high anesthetics such as propofol and also affect myogenic MEPs, but to a intravenous anesthesia using propofol and to be optimal and is the anesthetic regimen for monitoring of myogenic MEPs at but MEPs can be in most patients when of inhalational are used. have only effects on myogenic MEPs. produces changes in myogenic MEPs, but they can be recorded using a technique. of anesthetic on MEPs are described in in and The use of neuromuscular blocking during MEP monitoring Because changes in the anesthetic regimen may myogenic MEPs, when myogenic MEPs are to be the anesthetic regimen should be as as This is especially the time of such as of the spinal during spinal deformity surgery, or of a patient with cervical spinal or a bolus of anesthetic should be those Concurrent Monitoring of SEPs The use of MEP monitoring does not the for SEP monitoring of the spinal cord and The dorsal columns of the spinal cord and the somatosensory pathways in the brain may be during surgery without compromise of the corticospinal tracts by a spinal infarction or by a and without MEP when MEPs are used for intraoperative SEPs should be as Concurrent monitoring of SEPs and MEPs can detect compromise of either or motor tracts and also provides a of so that at method for monitoring the integrity of the nervous system is the other or to neurologic anesthetic or other technical and 2002). in the brain with TES are far lower than levels that have been demonstrated to be safe (MacDonald, during direct brain stimulation, but the high current can cause of the muscles and which in can cause In series of over with TES-MEP monitoring there were or and has also been (MacDonald and 2008). or should be used to or or damage during to of and could also risks to the Partial NMB may but may also complicate interpretation of the MEPs in some used, partial NMB should be done with a continuous infusion of the paralytic using such as the assessment of the responses to stimulation (Sloan and 2008) to assess the of NMB and the infusion If in the area of the surgical is large enough to interfere with the surgery, the timing of TES should be with surgical to avoid at times when this would be stimulation of the brain can This is a with direct cortical stimulation, especially with trains of pulses (the but also can with This of clinical during TES is in series of over during which TES-MEP monitoring performed (MacDonald, 2002). The of but activity to the during direct cortical stimulation is as is their clinical has not been shown that a of a patient to during TES, and such a should not be as a to The of monitoring during TES-MEP recordings is it is used in some centers but not in anesthesia during TES should be to should they If a the risk of a must be the of TES-MEP monitoring in to the central motor pathways in to TES-MEP monitoring. In the of MEP a of were considered to be relative to TES, a cerebral devices such as a or a and or under or to the stimulating electrodes (Legatt, 2002; 2002). However, in some or all of these are not as TES, and patients with these have TES-MEP monitoring (MacDonald and 2008). The of MEP monitoring must be the potential risks in each the of the changes in the IONM data should be rapidly to the of the surgical If the MEPs are not to neurologic compromise in the anesthesia, or technical this should also be to the they with surgical procedures in the that the MEP data are that the motor pathways are This is as an of the American Clinical Neurophysiology Society is on an assessment of current and clinical is not to include all possible of for a or all for to use a Neither is it to that patient are the of the patient and the for the on all The clinical section is made available to the with current and recommendations are not to clinical