Improving intraoperative evoked potentials at short latency by a novel neuro-stimulation technology with delayed return discharge

OBJECTIVE
The intraoperative monitoring of cranial nerve function records evoked responses at latencies of a few milliseconds. Unfortunately, these responses may be masked by the electrical artifact of the stimulation pulse. In electrical stimulation, the return discharge of the stimulation pulse significantly contributes to the width of the electrical artifact.


METHODS
We have generated stimulation pulses with an ISIS Neurostimulator (inomed Medizintechnik GmbH) providing a novel stimulation artifact reduction technique. It delays the return discharge of the stimulating pulse beyond the latency of the expected physiological response. This delayed return discharge is controlled such that no unintended physiological response is evoked.


RESULTS
In 21 neurosurgical interventions with motor evoked potentials of the facial nerve (FNMEP), the stimulation method generated a stimulation pulse artifact with reduced tail duration. Compared to conventional stimulation with immediate return discharge, the signal-to-noise ratio of the physiological response may improve with the novel stimulation method. In some surgeries, only the novel stimulation method generated clearly identifiable response signals.


CONCLUSIONS
The reduced width of the stimulation artifact extends the toolbox of intraoperative monitoring modalities by rendering the interpretation of cranial nerve evoked potentials more reliable.


SIGNIFICANCE
The novel technique enhances the number of patients for whom intraoperative monitoring may aid in cranial neurosurgery.


Introduction
Microneurosurgery of the skull base carries a significant risk of impairing cranial nerve function (Yas ßargil, 1984). Among the technical measures to preserve cranial nerve function, intraoperative neurophysiological monitoring (IONM) has become mandatory (Acioly et al., 2013;Deletis and Fernandez-Conejero, 2016). During surgery, IONM serves to communicate impending nerve damage to the surgeon and to predict the postoperative neurological state.
For continuous monitoring of cranial nerve motor function, transcranial electrical stimulation (TES) allows activation of the motor cortex and the motor pathway proximal to the surgical field and, ultimately, the recording of the motor evoked potential (MEP) in cranial nerve target muscles, for example the facial nerve MEP (FNMEP) (Akagami et al., 2005;Sarnthein et al., 2013;Bozinov et al., 2015;Seidel et al., 2020).
Despite these advantages, cranial nerve MEP monitoring has not become a standard tool of IONM yet. One reason is the close proximity between stimulation and recording sites that results in a large stimulation artifact and a short latency of the response. With standard stimulation techniques, the TES stimulation artifact may well extend over several ms after the stimulation pulse. The superposition of the artifact on the physiological response may render the interpretation of the results uncertain and thereby compromise IONM of cranial nerves.
As a novel approach, we present here a novel stimulation technique that drastically reduces the width of the simulation artifact. Among MEP of cranial nerves, the facial nerve is of highest interest -with examples of FNMEP we show that this technical advance is indeed relevant for intraoperative monitoring in cranial neurosurgery.

Patients
We included 21 patients (12 male, age 49 ± 21 y) who underwent neurosurgery at our institution (15 tumor and 6 vascular indications). Facial nerve function was at risk and FNMEP monitoring was performed. The collection of personal patient data and their analysis were approved and performed in accordance with the guidelines and regulations of the local research ethics committee (Kantonale Ethikkommission PB-2017-00094).

Facial nerve motor evoked potentials (FNMEP)
Intraoperative neurophysiological monitoring of the facial nerve was performed using the ISIS system (inomed Medizintechnik GmbH). Transcranial electrical stimulation was delivered by corkscrew electrodes placed at electrode sites C3/C4 versus Cz ( Fig. 1). We chose Cz as stimulation reference to ascertain the selective hemisphere stimulation. A bite block was placed in the mouth to prevent bite injuries of the tongue resulting from activation of jaw muscles. Transcranial electrical stimulation was performed by applying anodal rectangular pulses with a constant current stimulator. We recorded the responses from facial nerve target muscles orbicularis oculi, nasalis, orbicularis oris, or mentalis with 20 mm straight needle electrodes. The ground electrode (GND) was placed half-way between the stimulation and the recording sites. Responses were amplified and filtered (250-2000 Hz) before display (Fig. 2a).

Stimulation with the controlled delayed return discharge
In the following, we explain the difference between the standard stimulation method and the novel stimulation method with the controlled delayed return discharge. In electrical stimulation circuits, especially in the unipolar stimulation mode, there is a risk of electrolysis. In medical stimulation devices, this electrolysis is prevented by charge balancing through DC polarity compensation after stimulation. In standard devices, the charge balancing is performed by a coupling capacitor where the capacitor is discharged immediately after the stimulation pulse (Fig. 2b). This return discharge significantly contributes to the electrical stimulation artifact in the recorded signal, superimposes the physiological response and thereby renders it difficult to interpret the recorded signal (Fig. 2a). Our novel stimulation technique prevents this superposition. Instead of an immediate return discharge, we delay the return discharge of the stimulating pulse beyond the latency of the expected physiological response (Fig. 2c). This is achieved by a temporal cut-off of the current flow immediately after the stimulation. As a side effect, the frequency of the hardware filters for the recording channels needs to be lowered to 0.5 Hz. The delayed return discharge is controlled such that no unintended physiological response is evoked. This technique considerably reduces the influence of the stimulation artifact on the physiological signal of interest.
The novel stimulation method is covered by a patent (Baag et al., 2019). The safety of the device is assured by the CE (Conformité Européenne).

Facial nerve motor evoked potential (FNMEP)
To illustrate how the novel stimulation method improves FNMEP recording, we first show one example trace of a response in facial nerve target muscles from a FNMEP elicited by the standard stimulation technique (Fig. 2a). A train of 3 pulses (pulse intensity 80 mA, pulse width 0.5 ms) evoked a response in the facial muscles. The current flow of the standard stimulation technique is depicted schematically in Fig. 2b. A single control pulse delivered 40 ms before did not evoke a response (data not shown for clarity). The responses in Fig. 2a show a variety of latencies and are polyphasic to a varying degree. Common to all responses is the large transient of the stimulation artifact that is superimposed on the physiological response. While the response in the orbicularis oris and mentalis muscles can be clearly discerned, the response in the orbicularis oculi muscle is difficult to identify. The current flow of the novel stimulation technique is depicted schematically in Fig. 2c: the return discharge current flows only after a delay of 50 ms and with a flattened time course. Therefore the transient of the stimulation artefact as can be seen in Fig. 2d: the FNMEP muscle responses are clearly more distinguishable from the stimulation artifact and easier to interpret. The improvement is most striking here for the orbicularis oculi muscle.
To give an overview over the usefulness of the novel stimulation method with delayed discharge, we have collected data during surgery in 21 patients (Fig. 3). For better comparison, we show the FNMEP response of the orbicularis oris muscle only. Compared to the standard stimulation technique (grey lines) with the wider stimulation artefact, the novel stimulation technique with the narrower stimulation artefact (black lines) renders the interpretation of the curves more reliable.

Recordings
Given the large amplitude of the stimulation artifact and the small amplitude of the signal, FNMEP responses could not always be identified unambiguously with conventional stimulation. An The return discharge after the stimulation pulse is superimposed on the muscle responses. For small and early portions of the responses, interpretation is difficult. b) Current flow during conventional stimulation with a single pulse followed by a train of three pulses. Since all pulses are anodal there is a risk of electrolysis due to DC polarization. Electrolysis is prevented by DC polarity compensation immediately after stimulation. This charge balancing uses a coupling capacitor. The passive discharge of the coupling capacitor causes a current spike with opposite polarity to the stimulus and exponential decay of the charge. The resulting discharge curve depends on the intensity and duration of the stimulus and the capacitor used. c) Current flow with delayed return discharge. The goal of the novel stimulation method is to obtain no superposition of the charge balance on the response of interest. The method exploits that the response has limited duration after stimulation. The goal is achieved by delaying the discharge curve until after the duration of the response (orange arrow). Thus, the DC polarization is prevented without the negative effect of the immediate discharge superimposed on the response of interest. d) Stimulation of the FNMEP with delayed return discharge. The stimulation artifact extends less than 1 ms after the stimulation pulse. Therefore, the early small components of the relevant stimulus response signals are clearly visible. In particular, the response in the M. orbicularis oculi becomes amenable to interpretation. important cause for this ambiguity was the width of the stimulation artifact with the standard stimulation technique. This was especially striking in surgeries where the stimulation intensity had to be increased and the stimulus artifact presented with a larger amplitude. When delaying the return discharge with the novel technique, the situation signal-to-noise ratio improved in some surgeries as illustrated in Figs. 2, 3.
The novel stimulation technique improved the width of the stimulation artefact also in the muscle potentials that were evoked in other cranial nerve target muscles. Since facial nerve monitoring is most often required in the surgeries affecting cranial nerves in our institution, we focused here on the FNMEP.

Other methods to improve signal-to-noise ratio in FNMEP
The most obvious method to reduce the stimulation artefact in FNMEP is to reduce the stimulation intensity. However, the required stimulation intensity depends on several factors, among them the choice of sites for stimulation electrodes. Please note that the novel stimulation method with delayed return discharge has no effect on the required stimulation intensity.
The transient of the stimulation artefact has a slow decay compared to the muscle response (Figs. 2, 3). The frequency for the high-pass filter must therefore be chosen adequately, usually at 250 Hz. However, the filter may also diminish the muscle response. The delayed discharge reduces the width of the stimulation artefact at its origin without further constraining filters, which constitutes a conceptual advantage.
Another method to mitigate the effect of the wide stimulation artefact would be to alternate between anodal and cathodal stimulation pulses, which is a standard technique in electrophysiology. When averaging over several alternating stimulations to obtain the evoked potential, the stimulation artifact will be minimized in the average. However, to elicit FNMEP, anodal stimulation is required because when stimulating motor cortex, anodal stimulation elicits FNMEP responses at much lower thresholds than cathodal stimulation. Therefore, we always apply anodal stimulation and, consequently, charge balancing always requires cathodal current. Since alternating stimulation is not possible for FNMEP, the delayed return discharge is of particular advantage.

Limitations and strengths
We have presented here a novel technique in the light of 21 examples, which is not a clinical study. For Figs. 2 and 3, we have selected examples to illustrate the advantages of the delayed return discharge. In daily practice, intraoperative recordings may vary considerably depending on several factors. Clearly, even without delayed return discharge it is possible to obtain meaningful FNMEP (Akagami et al., 2005;Acioly et al., 2013;Sarnthein et al., 2013;Bozinov et al., 2015;Deletis and Fernandez-Conejero, 2016;Seidel et al., 2020). Conversely, in spite of the delayed return discharge, meaningful results may not be achieved in some surgeries. Furthermore, as a limitation of the novel technique, lowering the frequency of the hardware filters for the recording channels to 0.5 Hz may change the shape of the motor evoked potentials also in other recording windows. In our current practice, we weigh the advantages and disadvantages of the novel method in each surgery. Still, in our experience, we achieved meaningful results more often with the delayed return discharge.

Conclusions
The technique presented here improves the signal-to-noise ratio of evoked potentials with short latency. The reduced width of the stimulation artifact extends the toolbox of intraoperative monitoring by rendering the interpretation of cranial nerve evoked potentials more reliable. We have presented examples for this technical advance that opens the monitorability of short latency potentials for a larger group of patients. This increases the number of patients for whom intraoperative monitoring may aid in cranial neurosurgery.

Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: [M.T. and M.B are employees of inomed Medizintechnik GmbH. J.S. is member of the inomed scientific advisory board and has received speaker fees for giving presentations at inomed inhouse courses on IONM.]