Central nervous system pathophysiology also presents challenges to SSEP monitoring. Strokes, central demyelinating disorders such as multiple sclerosis, and other central disorders may have varying effects on SSEPs depending on the degree of central impairment. Central conduction delays, poor evoked waveforms, or absent evoked waveforms may result. In patients with demyelinating disease (e.g., multiple sclerosis), temperature management becomes particularly important as increased temperatures may impair neural transmission and therefore degrade signals in these patients (Uhthoff’s phenomenon). As such, it is important to document core temperatures and avoid relative hyperthermia in these patients.

Because scalp recording sites for SSEPs are a reflection of the electrode position relative to the brain tissue, any factor that alters this relationship may alter recordings. Scalp hematomas may dampen cortical response amplitudes; especially in patients with bleeding dyscrasia, a hematoma may even result from the placement of SSEP scalp recording needles. Similarly, subdural air, subdural hematomas/hygromas, or other tumors in this location have an even more profound effect and some of these may not be known to the intraoperative team at the time of surgery. When SSEP recordings point to dampening at a specific recording electrode rather than at all recording sites, one should consider a process possibly being present between the brain and scalp in addition to problems with the electrode itself. Similarly, a shift in the brain may occur after resection of a large tumor or when changing the position of the patient (supine to prone, supine to sitting position, or the reverse). Whereas the SSEP recording electrodes may stay in exactly the same position on the scalp, the underlying brain may shift slightly and therefore the relative orientation between brain and scalp may be altered and thereby affect SSEP signals.

During intracranial procedures, sterile field requirements may alter standard SSEP needle placements and thus changes in morphologies should be expected depending on the degree of displacement from standard electrode sites. In some cases, this alteration may be mitigated if the surgeon places sterile needles in the field. Neuronal migration disorders of cortical tissue (abnormal fetal migration of developing brain) or normal variation may also produce atypical SSEP cortical responses. For example, some patients will present with cortical waveforms following lower extremity SSEP stimulation that have a wider than typical N37 [1] distribution and a small or absent P37, thereby mimicking an upper extremity type of scalp distribution (“N37” morphology). Although this may be identified by its pattern if sufficient recording channels are utilized, verifying that the stimulation and recording electrodes are properly placed is also usually warranted for most atypical patterns (refer also to the section on technical issues in this chapter).

Patient pathophysiology that affects SSEPs may similarly affect transcranial electrical MEPs . However, given the difference and variability in the central neural pathways that mediate these monitoring modalities, there is a complementary nature between SSEPs and MEPs and one may remain more reliable than the other in the face of preoperative neurologic dysfunction .

Electromyography (EMG ) monitoring and pedicle screw stimulation testing is reliant on an intact motor unit, and anything that disrupts the motoneuron, root, nerve, neuromuscular junction, or muscle may have an effect. The presence of applicable disease states and particularly weakness in target muscles should warn the neurophysiology team of possible limitations in the use of these modalities. (For more information on nerve conduction studies, see Chap. 38, “Surgery in the Peripheral Nervous System”).

Radiculopathy or neuropathy may greatly reduce the sensitivity of EMG monitoring as mechanical perturbation of roots or nerves may not be as readily reflected at the muscle in these conditions. The location and duration of the lesion are influential in how these effects are demonstrated. An acute lesion leaves the distal nerve functionally intact, which can be dangerously misleading if the nerve is stimulated distal to the lesion yielding a robust compound muscle action potential (CMAP) and giving the impression it is functionally intact (e.g., in facial nerve testing). Thus, testing as proximal as possible is preferred to assure functional integrity of the nerve. With more chronic axonal lesions, Wallerian degeneration proceeds within a week and the distal nerve will also show dysfunction as well, although in rare cases the distal nerve may remain functional in a nonfunctioning nerve (neurapraxic conduction block).

Radiculopathy/neuropathy may also significantly impact testing that relies on stimulation thresholds such as pedicle screw electrical testing. In the normal setting, a high stimulation threshold suggests that the screw is well-insulated by bone and thus is appropriately positioned. However, falsely elevated thresholds due to nerve dysfunction may occur during pedicle screw testing in the presence of radiculopathy/neuropathy and, as such, test results involving weak myotomes should be interpreted with caution. If possible, direct proximal root stimulation may help determine if a root has a stimulation threshold in the normal range (e.g., <4 mA using typical stimulation parameters). It may also be helpful to examine relative thresholds if multiple screws are evaluated and the finding of a threshold that is much lower than companion screws should alert suspicion and receive additional evaluation for malpositioning from the surgeon. Similarly, for free run EMG monitoring, it should be kept in mind that sensitivity may be affected depending on the muscles tested and severity/distribution of the radiculopathy/neuropathy.

Pre-existing neuromuscular junction disorders may also influence EMG testing. Depending on the type of disorder, measures of neuromuscular blockade may also yield unreliable results and the response to pharmacologic neuromuscular blockade may be highly exaggerated and prolonged.

Electroencephalography (EEG ) is typically monitored either by scalp electrodes and/or by direct cortical recordings (electrocorticography) during intraoperative neuromonitoring. EEG may be useful for such intracranial surgeries as aneurysmal or other vascular repair, brain mass excisions, and in epileptic foci excision. It is also useful in extracranial surgeries like carotid endarterectomies to assess for hemispheric ischemia.

Unprocessed EEG is also useful in identifying the deep anesthetic state that exhibits a burst-suppression pattern (Fig. 17.3). This state is easy to identify and may be useful as a surrogate marker of other synaptic suppression affecting neuromonitoring signals (e.g., those related to MEPs) and the deep cortical anesthetic state supplies information to the anesthesia team that they may use to titrate anesthetics. On the other hand, if the goal is to assess through the entire range of anesthetic depth, EEG requires digital processing techniques (e.g., Bispectral Index [BIS] monitor ) that are often imperfect and fall outside the usual realm of the neuromonitoring team (see Chap. 11, “Clinical Application of Raw and Processed EEG”).

Those anatomic barriers that insulate or alter SSEP recording electrodes (as discussed previously) can have the same inhibitory effect on obtaining EEG traces. In particular, hygromas, subdural hematomas, and scalp hematomas are examples of fluid spaces that will insulate scalp needles from underlying cerebral generators of EEG. These and other preoperative abnormalities such as focal attenuation, epileptiform discharges, or rhythmic slowing are likely to be present prior to surgery, making preoperative or preincision EEG a helpful reference to distinguish between these pre-existing displays and new findings (Fig. 17.4).

The goals of anesthesia, analgesia, and amnesia may at times be in direct opposition to obtaining robust neurophysiologic data . Those neurophysiologic functions requiring more synapses are in turn typically the first to be blunted by anesthesia (e.g., conscious acts) and differential effects are noted within neurologic pathways utilized by common neuromonitoring tests (e.g., MEPs affected to higher degree than SSEPs). Although a bell curve may generally exist for the average anesthetic affect across patients, there is wide variation for an individual patient’s responses.

A number of systemic variables may also impact neuromonitoring data both directly and through the patient’s reaction to an anesthetic. Temperature, pH, blood pressure, hematocrit, oxygenation, CO ₂ levels, blood loss, fluid status, and other factors can impact monitoring signals, which can dramatically change intraoperatively [2]. Stability of anesthetic and systemic parameters improves the stability of neuromonitoring data and allows for a clearer interpretation. However, the reality is that some degree of variation is inevitable and wide degrees of change may be unavoidable. Understanding the impact that anesthetic/systemic variations have on neuromonitoring data and the ability to differentiate this from new neural dysfunction is part of the art of intraoperative neuromonitoring. For more details, see Chap. 19, “Anesthesia Management and Intraoperative Electrophysiological Monitoring.”

“Technical issues” is a broad term that concerns the technique used to acquire signals. As noted in the Introduction, these issues have their roots in good basic technique, and a solid working knowledge of neuromonitoring is assumed. Beyond basic technique, we hope to address further coping strategies in response to the patient and anesthesia challenges discussed previously, the sometimes electrically hostile operating room environment, equipment failure, and unintended deviations from usual methods.

As in all areas of signal optimization, a rational step-wise approach to understanding, localizing, and addressing the specific issues at hand will provide the best route to timely and effective optimization. In contrast, inexperienced persons may try steps “because they worked before” (often for a different problem) or in an otherwise illogical way. A moment of thought in these situations will often save valuable time, allowing redirection of attention to other optimization or even the monitoring itself. One extreme example is someone who wanted to “replace all the electrodes” when encountering a high-amplitude noise problem, but instead, with some quick investigation, showed that this was an external noise source (unplugging the bed solved the problem). Once issues are identified as technically related (as opposed to or in addition to patient/anesthesia-related issues), our goal is to further break down the issues until we come to a specific answer (Fig. 17.6).

For technical issues, we first decide whether the problem fits into one of three broad areas: increased electrical noise, poor signal amplitude, or atypical patterns of signal response. One will quickly recognize that the first two categories are the two components of the SNR, and deciding the root cause of a poor signal is usually deciding whether the problem is primarily increased noise, decreased signal, or both. This distinction hinges on both an understanding of the usual levels of background noise and the usual response characteristics of target signals. If we encounter a poor SNR and the amplitude of electrical noise exceeds our expectation for typically sized signals, then we’d expect that electrical noise is our primary issue. If, on the other hand, the level of electrical noise is below the range of typical signals, then we expect that low signal is the primary cause. This may appear trivial, but some common errors demonstrate that this thought process is not universally applied. For example, most of us have probably encountered a situation in which the sensitivity for recording has been increased to a point that universally present intrinsic noise is displayed prominently and the absent signals are described as “noisy.” Conversely, with high noise levels, recording sensitivity may be decreased to make the baseline superficially appear to have a reasonable noise level when, in fact, embedded signals would never be recognizable. The key to avoiding these errors is to have an understanding of expected amplitude and latency of the target signal. If one knows their target values, they will adjust their settings to those values rather than to a potentially deceiving “look” and the differentiation between noise and signal issues should be much easier.

The third main category for technical issues primarily deals with errors in the configuration of our monitoring systems or electrodes. This may be due to failure to correctly configure or connect components or it may be due to breakage of the equipment. When these “errors” occur, signals may be present in unexpected patterns/morphology, but possibly with reasonable recording fidelity. At other times, select signals will be absent in patterns that do not make physiologic sense or signals may be completely and unexpectedly absent but possibly with clues such as absence of stimulation artifact.

As is apparent, there is overlap between the “poor signal” category and the signal acquisition error category. In fact, one might logically consider this third category as a subdivision of the poor signal category. For discussion and thought organization purposes regarding the signal acquisition error category, we will focus on frank errors or breakage, whereas in the prior category, we will focus on reasonable but potentially suboptimal aspects of either the stimulation or the recording systems.