INTRAOPERATIVE NEUROPHYSIOLOGIC MONITORING

The main purpose of intraoperative neurophysiologic monitoring is to reduce postoperative neurological deficits, but more recently it has become apparent that intraoperative recording of sensory evoked potentials and electromyographic (EMG) potentials can also aid the surgeon during many operations. The use of intraoperative monitoring of sensory evoked potentials and EMG potentials to reduce permanent postoperative deficits is based on the assumption that changes in recordable electrical responses occur as a result of injury, and that the injury is still reversible at the time of detection if proper surgical intervention occurs.

Monitoring of brain stem auditory evoked potentials (BAEPs) during operations in which the vestibulocochlear nerve (cranial nerve VIII) may be manipulated is now widespread, and the use of monitoring to reduce the incidence of hearing loss due to surgical manipulation of the vestibulocochlear nerve is steadily increasing. However, intraoperative monitoring of visual evoked potentials (VEPS), has not gained similar acceptance, mainly because of the technical problems involved in generating a suitable stimulus. Intraoperative recording of somatosensory evoked potentials (SSEPs) is valuable during aneurysm surgery and during other vascular operations, as well as during the removal of tumors in which the tumor itself or the surgical manipulation used to remove the tumor can be expected to affect brain structures that are involved in the somatosensory system and thus affect the generation of SSEPs.

The most frequent use of intraoperative monitoring of SSEPs is in operations involving the spinal cord, in which such monitoring is now well established. More recently, intraoperative transcranial stimulation of the motor cortex, by using either a high-voltage electrical stimulation or a strong magnetic field impulse in connection with recording of EMG potentials from the motor system, has been introduced to reduce intraoperative injuries to the spinal cord.

The recording of EMG responses from muscles innervated by different cranial nerves has been shown to be valuable in identifying cranial motor nerves during operations to remove tumors when the anatomy has been altered by the tumor or by previous operations; this is particularly true for operations to remove acoustic tumors and it has more recently been shown to be of value also in operations to remove large tumors of the skull base. It has been possible in a few types of operations to use electrophysiologic methods to guide the surgeon in the operation and to ensure before the operation has ended that the therapeutic goal of the operation has been achieved.

The objective of intraoperative monitoring of evoked potentials (BAEPs, SSEPs. and VEPs) when used for the purpose of reducing permanent postoperative neurological deficits is to detect changes that occur during the operation. This differs from the goal of the use of evoked potentials for diagnostic purposes, in which a deviation from a normal value is of interest. In intraoperative monitoring, normal values are of little interest; instead, it is important to obtain a baseline recording from an individual patient and then to compare the potentials that are recorded during the operation to that baseline recording. Such a baseline can usually be obtained after the patient has been anesthetized but before the operation has begun.

Monitoring of contractions of the facial muscles is performed during operations in the ccrebellopontine angle to help the surgeon locate the facial nerve (CN VII) when it is not identifiable visually. Possibly just as important, this technique makes it possible to determine which portions of an acoustic tumor do not contain any part of the facial nerve. This thus allows the surgeon to remove portions of a tumor without risk of injuring the facial nerve. Currently most such monitoring involves the surgeon using a handheld stimulating electrode, which carries short pulses of electrical current, to probe the surgical field to identify the facial nerve. Various methods are used to record the subsequent contractions of the facial musculature. The facial muscle contractions that are elicited by irritation and manipulation of the facial nerve are just as important as the contractions elicited by electrical stimulation of the facial nerve for the purpose of assessing injuries to the facial nerve.

Earlier, it was customary to have an assistant observe any movement of the face and then to communicate that fact to the surgeon. More recently, the recording of EMG potentials from the facial musculature and the recording of movements of the face using electronic sensors have also been used to verity that the facial nerve has been stimulated. The use of EMG recordings makes it possible to assess the degree of facial muscle contraction quantitatively, which was not possible when facial muscle activity was assessed by visual observation of movements of the face. The capability of making the recorded EMG potentials audible allows the surgeon to hear when the facial nerve has been stimulated and thus the surgeon has no need to rely on communication with an assistant. For this purpose, only a conventional audio amplifier and loudspeakers need to be connected to the output of the EMG amplifier. Some commercial equipment makes use of amplitude-sensing devices that elicit a tone signal when the EMG potentials reach a preset amplitude. However, making the (original) EMG signal audible is advantageous because it provides valuable information that is lost when such amplitudesensing devices are used. Although the audible EMG signal provides most of the information that is needed when monitoring the facial nerve, an oscillographic display of the recorded EMG potentials is also advantageous, in that it allows assessment of the amplitudes and latencies of the recorded EMG potentials. When the movements of the face are recorded by electronic sensors, the movements can be made to elicit a sound (horn), but an oscillographic display of the electrical signals produced by movement detectors has limited value.

Because the facial nerve is often split into several fascicles when a large acoustic tumor has displaced it, it is important that recordings be made from the entire face. If only portions of the facial musculature (e.g., the lower face or the upper face) are monitored, failure to locate all parts of the facial nerve could result in inadvertent removal of or injury to portions of the facial nerve: this would result in postoperative paralysis of part of the face. Usually the record EMG activity is performed between two electrodes, one placed on the forehead and one on the lower face (Fig.-1A), when monitoring facial function intraoperatively. In addition to recording activity from all muscles on the side of the face on which the electrodes have been placed, this particular arrangement of recording electrodes will also record contractions of the masseter and temporal muscles. These muscles are innervated by the motor portion of the trigeminal nerve (portio minor), and there is the possible risk of mistaking the fifth (motor) nerve for the seventh nerve when probing the cerebellopontine angle for the facial nerve in operations to remove large acoustic tumors that have progressed rostral to the trigeminal nerve. The EMG response to stimulation of the trigeminal nerve, recorded in the way illustrated in Fig-1A, however, can easily be distinguished from the response to stimulation of the facial nerve, because the latency of the recorded EMG signal differs in the two situations [1.5 ms and 5 to 6 ms. respectively,(Fig-1B). An alternative way to distinguish between the response to facial nerve stimulation and that to stimulation of the motor portion of the trigeminal nerve (CN V) is to record from the masseter muscle using a pair of electrodes connected to a separate differential EMG amplifier (Fig-1A). This recording will almost exclusively yield the response of the masseter muscle, and thus is a good indicator of stimulation of the motor portion of the trigeminal nerve.

Fig-1: A-Electrode placement for recording EMG potentials from facial muscles and a separate recording of the response from the masseter muscle. B- EMG potentials recorded from electrodes, placed as shown in A, elicited by electrical stimulation of the trigeminal and facial nerves in the CPA. C- Hand held stimulating electrodes for intracranial localization of the cranial nerves.

There are several advantages to using EMG recordings as a measure of muscle contraction rather than using a single sensor that records movements of the face. First, EMG potentials from practically all of the facial muscles can be recorded on a single channel [Fig-1A), whereas several sensors are needed to cover the entire face when movements are being recorded. Second. recording EMG potentials makes it possible to measure the latencies of the responses accurately, which enables one to differentiate between activation of the trigeminal nerve and activation of the facial nerve (Fig-1B). Third, the amplitude of the EMG response is roughly a measure of how many nerve fibers are functioning. and it therefore provides valuable quantitative information about the degree of injury to the facial nerve.

There are essentially two ways that electrical stimulation can be applied to the facial nerve: one is by using a bipolar stimulating electrode, and the other is by using a monopolar electrode. the difference being that the bipolar electrode has a higher degree of spatial specificity. However, because it is only the negative phase of the stimulus that is effective, only one of the two prongs of a bipolar stimulus electrode will stimulate the facial nerve effectively. Therefore, the orientation of a bipolar electrode is important. A monopolar. hand-held, stimulating electrode does not have these disadvantages and although it is less selective than a bipolar stimulating electrode it is preferable for intraoperative use (Fig-1C). A monopolar stimulating electrode is easy to use and its stimulating power is not affected by its orientation, as is the case for a bipolar stimulating electrode. By using a monopolar, handheld, stimulating electrode and having EMG potentials made audible, the surgeon can probe a large area of a tumor quickly and "map" the tumor to locate portions of the tumor where no part of the facial nerve is present so that it can be removed safely. The use of this technique can often reduce considerably the time required to remove the tumor because large portions of the tumor can be removed without risk of causing injury to the facial nerve. Later during removal of the tumor, when it becomes important to identify the facial nerve accurately so that injury to the nerve can be avoided, the same monopolar stimulator can be used and the area of tissue that it stimulates can be varied by varying the voltage that is applied through the electrode.

The electrical stimulation should consist of negative (rectangular) impulses of short duration (0.1 to 0.2 ms), and the stimulus strength should be no greater than necessary to produce a contraction. Some older types of stimulators make use of large current and some even make use of direct current. Such stimulators should not be used because of poor specificity and, particularly, because of the risk of injuring the nervous tissue with the electrical current used to stimulate the facial nerve.

Because shunting of electrical current from the stimulating probe can vary widely when the surgical field is wet compared to when it is relatively dry, it is advantageous to use a relatively constant-voltage stimulator rather than the more conventional constant-current type of stimulator. Constant-current stimulation in connection with the use of a "flush tip" stimulating electrode (i.e., an electrode that is insulated all the way to its tip) has been advocated by some investigators. When constant-voltage stimulation is used, the same amount of stimulus current will flow through a specific tissue (e.g., the facial nerve), regardless of how much shunting occurs. If constant-current stimulation is used, the same total current is delivered, but the amount of current that passes through a specific volume of tissue depends greatly on how much current is shunted away; in this case the stimulus strength depends heavily on whether the field is wet or dry.

When a facial nerve stimulator is used to identify regions of a tumor where no part of the facial nerve is present, the stimulus strength should be set so that it will activate the facial nerve if the nerve is within a small distance of the tip of a monopolar stimulating electrodes.

Audible monitoring (by means of a loudspeaker) of the EMG activity of the facial muscles that is evoked by mechanical stimulation of the facial nerve provides valuable feedback to the surgeon during the delicate resection of portions of an acoustic tumor that involve the facial nerve. Such continuous monitoring of facial EMG activity (without electrical stimulation) is of great value, particularly when the surgeon is removing a large tumor, parts of which may be firmly adherent to the facial nerve. If the facial nerve is being heated by electrocoagulation, or heated by the drilling of bone adjacent to the facial nerve, transient or sustained facial muscle activity will result. Although such monitoring of facial EMG activity is important in reducing the risk of permanent damage to the facial nerve, it must be pointed out that the facial nerve can be injured permanently from surgical manipulation without any EMG response being noted; thus, injury from sharp dissection will most likely not result in recordable EMG activity (or any movement of the face). For this reason, when the surgeon is dissecting near the facial nerve spontaneous EMG activity should not be relied upon for assurance that the facial nerve remains intact; in this situation, electrical stimulation of the facial nerve should be used frequently to identify the facial nerve so that the surgeon remains aware at all times of the exact location of the facial nerve.

In operations to remove tumors of the skull base, several cranial motor nerves are often in the operative field and are thus at risk of injury from surgical manipulation. This is particularly true in operations within the cavernous sinus, where the nerves that innervate the extraocular muscles [the oculomotor (CN III), trochlear (CN IV), and abducens (CN VI) nerves] may be difficult to identify visually or may be displaced by the tumor. Recording EMG potentials from the extraocular muscles while the surgical field is being probed with a hand-held stimulating electrode similar to that used for stimulating the facial nerve (Fig-1C) can aid the surgeon in locating the respective nerves. EMG potentials can easily be recorded from these muscles via needle electrodes inserted percutaneously into the respective muscle (Fig-2). The potentials recorded from the extraocular muscles in response to electrical stimulation of the respective nerves are easy to distinguish (Fig-3A).

More recently, surface electrodes for recording EMG activity from the extraocular muscles have been developed. Facial nerve function should also be monitored in these operations, using the methods just described. Continuous recording of EMG potentials from these muscles is also important, because injury to the respective nerves from mechanical manipulation and from heat during electrocoagulation will often result in transient or sustained EMG activity, as was described for the facial nerve. Thus, such activity can be an important aid in preserving these nerves.

Fig-2: A-Electrode placement for recording from facial, extraocular muscles and the tongue. The reference electrodes are all placed on the forehead on the opposite side, to avoid recording from the facial muscles at the same time. B-Electrode placement with intraoperative recordings from extraocular and facial muscles. BAEPs and VEPs included.

Fig-3: A- EMG potentials recorded from the extraocular and facial muscles in response to intracranial electrical stimulation of the respective cranial nerves, using the electrode arrangement as seen in Fig-2. B-EMG potentials from the tongue in response to stimulation of the hypoglossal nerve.

Identification of the hypoglossal nerve (CN XII) can be facilitated by recording EMG potentials from the tongue (Figs-2A and 3B), and monitoring of the accessory nerve (CN XI) can conveniently be done by placing pairs of needle EMG electrodes in the trapezius muscle (Fig-2A). The motor portions of the glossopharyngeal (CN IX) and vagus (CN X) nerves can be monitored by stimulating the respective nerves electrically and recording EMG activity from the muscles that these nerves innervate in a way similar to that just described for the facial nerve and the nerves that innervate the extraocular muscles. A pair of needle electrodes placed in the soft palate will record the EMG response to stimulation of the glossopharyngeal nerve and electrodes placed in the supraglottic region are suitable for recording of the EMG response from laryngeal muscles that are innervated by the recurrent nerve of the vagus nerve. Electrodes placed on the endotracheal tube can record surface EMG potentials from laryngeal muscles. A balloon placed on the tip of an endotracheal tube as a pressure recording device has been used to record contractions of the laryngeal muscles for monitoring of the recurrent laryngeal nerve. This technique could be useful in monitoring the more central portions of the vagus nerve.

One should be aware when monitoring the glossopharyngeal, vagus, and accessory nerves using electrical stimulation that there are risks involved. For example, a supramaxial stimulation of the accessory nerve may result in so strong a muscle contraction that dislocation of joints or physical injury to muscles and tendons may result. Electrical stimulation of the glossopharyngeal and vagus nerves may cause cardiovascular effects and should therefore be done cautiously.

Microvascular decompression to relieve hemifacial spasm is one of only a few operations in which intraoperative neurophysiologic monitoring can aid the surgeon in achieving the therapeutic goal of the operation. It is generally accepted that hemifacial spasm is caused by vascular compression of the facial nerve as it exits the brain stem. and that microvascular decompression of the root exit zone of the facial nerve is the most effective treatment of this disorder. However, it is not always obvious from an exploration of the root exit zone of the facial nerve which of several vessels is causing the spasm, and a certain (small) number of patients who have undergone this operation have experienced spasm postoperatively. Some of these patients had to be reoperated upon, depending on the severity of the spasm.

In studies of the pathophysiology of hemifacial spasm, it was found that an abnormal muscle response that seems to be characteristic of the disorder disappears instantaneously when the offending blood vessel is moved off the intracranial portion of the facial nerve (Fig-4). This abnormal muscle response, which has a latency of about 10 ms is seen when one branch of the facial nerve is stimulated electrically and recordings are made from muscles that are innervated by other branches of the facial nerve. By monitoring this abnormal muscle response intraoperatively, it is possible to identify the offending blood vessel and to ensure that the nerve has been fully decompressed by watching for the cessation of the abnormal response (Fig.4B) When using this method it was found that even veins can cause hemifacial spasm and that in many cases there was more than one blood vessel compressing the facial nerve. There are reasons to assume that at least some of the patients who experienced only partial relief from their symptoms before this type of monitoring was introduced did so because more than one vessel was affecting the facial nerve and only one of the offending vessels was moved off the facial nerve during the operation.

Intraoperative monitoring of BAEPs is commonly performed to reduce the risk of hearing loss as a result of intraoperative manipulation of the vestibulocochlear nerve in operations in the cerebellopontine angle. This is important in operations on acoustic tumors, in which hearing preservation is anticipated, as well as in microvascular decompression operations on cranial nerves and in other operations in the cerebellopontine angle. BAEPs are commonly recorded between electrodes placed on the vertex and on the earlobe (or mastoid) of the ear to which the sound is applied. BAEPs are best elicited by click sounds presented at a rate of 20 to 40 pulses/s (pps) at intensities of 100 to 110 peak equivalent sound pressure (Pe SPL), The normal BAEPs are characterized by 5 to 7 vertex-positive peaks that are generated as the different structures of the ascending auditory pathway are successively activated (Fig.-5). Because the BAEPs are generated by fiber tracts and nuclei of the ascending auditory pathway in the brain stem, the vestibulocochlear nerve, the cochlear nuclei, and the lateral lemniscus being the most important generators of BAEPS-recording of these potentials is not only useful for detecting injury to the vestibulocochlear nerve but may also be of value in operations in which the brain stem is being manipulated or when circulation to the brain stem may be compromised. The nuclei of the ascending auditory pathway are sensitive to manipulations of the brain stem, and there are indications that BAEPs may be more sensitive in detecting such effects than are changes in heart rate and blood pressure. A change in the latency of peak V while that of peak III remains unchanged (increased III - V interpeak latency) indicates an effect from surgical manipulation on structures located in the region of the superior olivary complex on either side or on the lateral lemniscus and its nucleus on the side contralateral to the side on which the BAEPs are being elicited (tumor side) (Fig-6). It is therefore more valuable to monitor BAEPs during other operations in which manipulation of the brain stem may occur, such as the removal of large acoustic tumors or other tumors of the cerebellopontine angle.
Fig-5: Neural generators of BAEPs. DCN: dorsal cochlear nucleus, VCN: ventral cochlear nucleus. SO: superior olivary complex. LL: lateral lemniscus. IC: inferior colliculus. MG: medial geniculate.

Fig-6: Changes of latencies and amplitudes of peaks III and V of BAEPs as a function of time during surgery. BAEPs were elicited by stimulating the opposite ear relative to the tumor location.

SSEPs are important in monitoring sensory conduction in the spinal cord. SSEPs elicited by stimulation of the median nerve and recorded from the contralateral parietal region of the scalp (C3 or C5) using a noncephalic reference are characterized by a series of positive and negative peaks (Fig-7). P₉, P₁₁, and P_I4 are short-latency, positive peaks that are generated at the level of the brachial plexus (P9). spinal entry (P11). and termination of the dorsal column in the dorsal column nuclei (P_I4). The bilateral N₁₈ is generated in brain stem nuclei, such as the superior colliculus. whereas the contralateral N₂₀ is assumed to be generated in the primary cortex. The waveform of the SSEPs and the presence of certain components depend on the recording sites (Fig-7). For monitoring the spinal cord, SSEPs are elicited by electrical stimulation of sensory nerves on the leg (peroneal or posterior tibial nerve).

Only when the cervical spine is considered is it appropriate to use SSEPs elicited by stimulation of the upper limb (median nerve at the wrist). The response to stimulation of the lower limb is recorded from electrodes placed on the scalp, vertex to a midline front reference or vertex to a noncephalic reference. The different components appear with longer latencies than those of SSEPs elicited from the median nerve. and the pattern of the peaks is more complex (Fig-8). One reason that lower limb SSEPs are more complex than those elicited from the upper limbs is related to the fact that ascending neural activity elicited by stimulating lower limbs travels in two separate fiber tracts in the spinal cord. Only part of the sensory information is relayed in the dorsal column nuclei (nucleus gracilis).

Fig-7: SSEPs in response to electrical stimulation of the median nerve. A- The thick line is a record from the contralateral parietal region with a noncephalic reference. B- Records from prerolandic region with a noncephalic reference.

Fig-8: SSEPs in response to stimulation of the left posterior tibial nerve at the ankle. A-Records from a midfrontal (pF_z). B- Record from midpaietal (P_z) with a noncephalic reference on the shoulder. C- Difference between A and B.

Neural activity that travels in the dorsal column is mainly elicited by skin receptors. Fast-conducting afferents that conduct activity elicited by stretch receptors, group Ia and II afferents from muscle spindles, and tendon organs travel in a spinothalamic pathway in the dorsal lateral funiculus of the spinal cord and synapse in the nucleus gracilis, after which they join other afferents in the medial lemniscus traveling toward the thalamic relay nuclei. The slow cutaneous afferents from the lower limbs that travel in the dorsal column are relayed in the dorsal column nuclei (gracilis nucleus). These afferents have a large range of conduction velocities, which causes a temporal dispersion in the elicited activity, which is the cause of the low amplitude of the early response from stimulation of lower limbs. This makes the SSEPs from lower limbs qualitatively different from the SSEPs elicited by stimulation of upper limbs and, together with the longer distance from the location of the stimulation to the brain stem structures, explains why the far field SSEP responses to electrical stimulation of lower limbs are less well defined than the SSEPs elicited by stimulation of the upper limbs, where all somatic afferents trawl in the dorsal column and are all relayed in the dorsal column nuclei (cuneate nucleus). Although recording using a noncephalic reference is appropriate for identifying the neural generators of SSEPs, the unfavorable signal-to-noise ratio in such recordings has made it more common to place the reference electrode on the scalp when SSEPs are used for intraoperative monitoring. This reduces the pick-up of electrical interference signals. When the spinal cord is to be monitored, all components of SSEPs that originate from structures that are rostral to the location where injury may occur can be utilized for detecting injuries. However, late components of the SSEPs are affected by anesthesia, and it would therefore be advantageous to use early components such as the P₁₄ in upper limb SSEPs. Unfortunately, the early components of lower limb SSEPs are less well-defined and usually cannot be used for intraoperative monitoring.

Because the blood supply to the portion of the spinal cord that comprises the ascending somatosensory pathway is different from that of the descending motor tracts, there is a possibility that the motor system can be injured without any noticeable change occurring in SSEPs. Thus, it seems possible that severe injury could go unnoticed when only SSEPs are monitored. Although it has been disputed whether this is in fact a real risk, there is a need to be able to monitor both motor and sensory systems during operations in which the spinal cord is at risk. Although traditional techniques can be used to monitor the somatosensory system intraoperatively, there are considerable technical obstacles in eliciting motor responses upon stimulation of the motor cortex. Therefore, such monitoring of the descending motor tracts is not done routinely. It has, however, been shown that it is indeed possible to elicit motor responses by electrical and magnetic transcranial stimulation of the cortex, thus a prerequisite for developing routine methods for intraoperative monitoring of motor tracts.

Intraoperative monitoring of SSEPs is valuable as an indicator of decreased cerebral perfusion in regions of the brain on which the generation of more long-latency components of the SSEPs depend. The development of these methods, which are now in routine use, had been pioneered by Symon et. al, who made use of recordings of SSEPs elicited by electrical stimulation of the median nerve at the wrist. The use of this method is based on the finding that there is a rather close relationship between the disappearance of electrical activity and a decrease in the cerebral blood flow to 15 to 18 ml/100 g per min (or below), and there are changes in the late components of the SSEPs that occur before the perfusion reaches these low levels. SSEPs are therefore valuable in estimating changes in cerebral blood flow.

The most useful parameter of SSEPs as an indicator of a decrease in cerebral blood flow is the central conduction time, which is the time difference between the occurrence of the components of the SSEPs that can be recorded at the neck (P₁₄) and the N₂₀ component that is recorded at the contralateral parietal skull (C3-C5) (Fig-9). Central conduction time is not affected by changes in peripheral neural conduction or by age, and it has been shown to correlate with cerebral blood flow when blood flow is reduced below a certain value.

Fig-9: Measures of central conduction time from recordings of the SSEPs that are elicited by electrical stimulation of the median nerve at the wrist. All recordings were done with a noncephalic reference. A- Parietal scalp. B- Frontal scalp. C- Spinal C6 spine.

During operations in which the optic nerve or the optic tract is being manipulated it would seem to be beneficial to monitor VEPS. However, the results of such monitoring have been generally disappointing because the changes in the recorded potentials correlate poorly with postoperative changes in vision. This is most likely due to limitations inherent in the techniques currently available for intraoperative stimulation of the visual system, and perhaps to inadequate knowledge of how to interpret the changes in the VEPs that may occur during surgical manipulations of the optic nerve or optic tract: however, essentially there are no practical problems involved in recording VEPs intraoperatively. At present, the only practical type of visual stimulation that can be used intraoperatively is the "flash" type. It has been shown that VEPs elicited by a changing pattern (pattern reversal-checkerboard pattern) are much more useful diagnostically than VEPs recorded in response to flash stimulation, however, eliciting VEPs by pattern reversal technique requires that a pattern be focused on the retina, which is not possible to accomplish intraoperatively.

Although recordings of BAEPs are not noticeably affected by any common anesthetic regimen using inhalation anesthetics. barbiturates. and other intravenous anesthetics, intraoperative monitoring of cranial motor nerves cannot be done if the patient is paralyzed, because such monitoring depends on recording muscle responses by EMG technique. Components of the upper limb SSEPs that occur with latencies longer than 16 ms are usually affected by general surgical anesthesia.

The most commonly used anesthetic regimen for neurosurgical operations involves giving a strong narcotic, such as fentanyl for analgesia together with an inhalation agent, such as nitrous oxide, and small amounts of halogenated agents ("balanced anesthesia"). Because the patient must be kept from moving, he or she must be paralyzed when such a regimen is used. Normally some form of a muscle endplate blocker (pancuronium, vecuronium. etc.) is used for this purpose. Therefore, balanced anesthesia makes it impossible to record EMG potentials because of muscle relaxation. However, intraoperative neurophysiologic monitoring is fast becoming a regular component of the surgical arena, and it is now common to adjust the anesthesia regimen to the requirements of the specific type of monitoring to be done. For instance, the requirement of being able to record EMG potentials has resulted in the use of inhalation anesthesia throughout the operation. possibly with the addition of a small amount of a narcotic agent and benzodiazepines such as midazolam. This type of anesthesia has been used for many years without any noticeable difficulties or complications and is now in common use when intraoperative monitoring of muscle activity is performed. However, while such an anesthesia regimen has no noticeable effect on recording EMG potentials or BAEPs and short-latency SSEPs. it is likely to suppress later components of SSEPs and it greatly suppresses motor evoked potentials. Some anesthetic agents such a propofol, that can be administered intravenously have less suppressive effect on cortical responses and are thus being used during operations in which several different kinds of evoked potentials and EMG responses are to be monitored.

Patients in whom intraoperative neurophysiologic monitoring of evoked potentials is to be performed should have the systems that are to be monitored quantitatively evaluated preoperatively. For example, if a patient's BAEPs are to be monitored, his or her hearing status should be evaluated preoperatively. Testing should include at least a pure tone audiogram and a determination of speech discrimination scores, and the patient's ear canals should be checked for obstruction (from cerumen. etc.). If the patient has no hearing or if it is not possible to obtain an interpretable BAEP before the operation, it will not be possible to obtain one during the operation. If, on the other hand, an interpretable BAEP can be obtained before the operation but not intraoperatively. the cause is most likely technical: such problems then must be corrected. Similarly, if SSEPs are to be recorded intraoperatively, a recording should be performed preoperatively to ascertain that the patient has normal SSEPs or to establish the degree of abnormality so that it can be taken into consideration before the intraoperative recordings are made. Factors such as neuropathies that may affect neural conduction should be ruled out or assessed quantitatively, if possible. If the extraocular museles are to be monitored, their function should be evaluated preoperatively by state-of-the-art methods. When facial EMG responses are to be monitored, the function of the facial musculature should be evaluated preoperatively. Preoperative evaluation of sensory and motor systems that may be at risk during an operation is also essential, because this, in connection with similar tests done postoperatively, the only way that deficits from the operation can be assessed quantitatively.

Intraoperative neurophysiologic monitoring of evoked potentials has been introduced for use in neurosurgical operations for long time, and it is therefore natural that the method has been viewed with skepticism in the early period of its implementation, and suffered from a demand for proof of the benefits. As with several other procedures allied to such operations, it has generally been difficult to design studies to provide quantitative estimates of the benefits in terms of reduced postoperative complications. This is largely because many surgeons believe that intraoperative monitoring is valuable in reducing postoperative deficits and therefore will not allow random selection of patients to be monitored. This makes it impossible to study the efficacy of such techniques by a double-blind methodology.

Comparing complications before and after the introduction of intraoperative monitoring thus seems to be the only practical way to assess the value of intraoperative monitoring: however, the results of such evaluations are influenced by any changes in the operative technique introduced at or after the institution of monitoring. Despite this complication. comparisons have been made between the rate of complications (such as facial palsy) of surgical procedures (such as acoustic tumor removal) before and after the introduction of intracranial monitoring of cranial nerve function. Such studies have shown a significant decrease in complication rate after the introduction of intraoperative monitoring.

It is generally accepted now that intraoperative monitoring of facial nerve function during operations on acoustic tumors is of value in preserving the facial nerve. A study of a less frequently performed operation, that to relieve hemifacial spasm, showed that the efficacy of the operation was higher after intraoperative monitoring was introduced. Further, the number of patients who needed to be reoperated upon because of recurrent or unrelieved spasm decreased from about 15 percent to almost zero. Finally, the frequency of hearing loss as a complication of microvascular decompression operations in the cerebellopontine angle decreased radically after the introduction of intraoperative monitoring of BAEPs during such operations.

Because intraoperative neurophysiologic monitoring can many times identify exactly which step in an operation that caused an injury is likely to result in a permanent neurologic deficit, it has contributed to the development of better and safer operating techniques. When evaluating the benefits of intraoperative neurophysiologic monitoring it must be pointed out that the benefit from intraoperative neurophysiologic monitoring depends on the level of experience of the surgeon. Thus a very experienced surgeon will not have the same degree of benefit from this technique as might a less experienced surgeon. There are several advantages of the use of intraoperative neurophysiologic monitoring, but they are difficult to measure quantitatively; nevertheless, they are great enough to lead most surgeons who have operated with the aid of such monitoring to demand that it continue.