The introduction of intraoperative neurophysiological monitoring technique for the auditory function emerged in the 1970s and at the beginning of the 1980s together with the first attempts of hearing preservation in CPA surgery [32, 47].

The goal of intraoperative auditory function monitoring is to provide information about the auditory pathway status during surgery in order to prevent damage to the auditory function [48] and consequently hearing loss. Early detection of the damage through a near-real-time recording technique is required for the electrophysiological method to alert the surgical team, resulting in immediate interruption or change of the microsurgical maneuvers with the assumption for reversal of the damaging process. This may enable wave recovery sparing hearing loss [48, 49].

The brainstem auditory evoked potential (BAEP) is the most widely used intraoperative electrophysiological method for auditory function monitoring, because it has a high sensitivity [48] and reliability to detect cochlear nerve damage [50]. However, BAEP requires 30 s up to 4 min to deliver reliable response of wave changes [34, 49], which is a disadvantage of the method. In this interim, a permanent damage to the cochlear nerve could happen, preventing the surgical team from detecting and avoiding it [48].

Direct auditory nerve manipulation [49, 51, 52], cerebellar retraction [49, 51, 52], and drilling of the internal auditory canal [49] are identified events that lead to intraoperative BAEP changes and, ultimately, could intervene with the postoperative hearing outcome.

During intraoperative BAEP monitoring, three changes can happen, as follows: latency increase, amplitude decrease, and wave loss [49]. Usually, latency increase and amplitude decrease are observed in the way to wave loss [49].

“When to warn the surgeon?” still is a matter of debate between authors. Empirical and divergent warning criteria have been used for different electrophysiological teams to warn the surgeon about deteriorating BAEP: (a) latency increase of 0.3 ms or amplitude decrease of 50 % [49], (b) latency increase of wave V of 0.5 ms [48], (c) latency increase of wave V of 0.07 ms/min or more than 1.5 ms (absolute value) [51], or (d) latency increase of wave V of 1.0 ms and amplitude decrease of 50 % [48, 53, 54].

Latency increase of wave V of 1.0 ms and amplitude decrease of 50 % are the most widely used criteria [48, 53, 54]. However, the observance of hearing impairment in CPA tumor surgeries following increased latency of less than 1 ms together or not with amplitude decrease smaller than 50 % of wave V showed that these criteria may be not sensitive enough as “warning sign” for hearing preservation in this group of patients [54]. Conversely, in patients affected of nonneoplastic CPA diseases, such as neurovascular compression (trigeminal neuralgia, facial hemispasm, etc.), these criteria seemed to be too sensitive in the form that latency increase of wave V of 1.0 ms and amplitude decrease of 50 % or even wave V loss did not predict hearing loss in this group of patients [54].

So, the reasonable answer for that question would be: “It depends on the disease that you are dealing with” [53, 54]. For CPA tumor a more strict criteria should be used and, therefore, we use latency increase of more than 0.3 ms or amplitude decrease greater than 50 % as our reference in such surgeries [49].

As previously reported by Matthies and Samii [49], each I, III, or V temporary or permanent wave loss carries a risk of deafness, as follows: for wave I loss is 65 %, for wave III loss is 65 %, and for wave V loss is 78 %. Whether these losses are concomitant, that is, permanent waves I, III, and V loss or type B5, it is associated with 96 % of deafness postoperatively [49]. The loss of wave III is the most precocious and sensitive sign for predicting postoperative deafness, whereas loss of wave V is the most definitive sign of postoperative deafness [49]. Other authors also emphasize the relevance of waves I and V loss for hearing preservation, alone or in combination [51, 52].

BAEP stability during CPA tumor surgeries indicates high hearing preservation rate [52, 55]. Patients with preoperative types B1 and B2 have a higher rate of hearing preservation (up to 80 %), whereas those with types B4 and B5 have reduced chances of hearing preservation (12–17 %, respectively) [32].

BAEP is performed throughout the procedure at our department in all patients that have types B1 to B4 BAEPs being the microsurgical maneuvers modified if BAEP changes happen that are known to have impact on postoperative hearing.

Even though there is an increasing rate of reported facial nerve preservation following VS surgery [56], facial nerve paresis or paralysis remains a significant risk to the patients [57], being frequent postoperative sequelae of major concern [56]. Indeed, the increasing rates of facial nerve preservation are supported by the introduction of routine intraoperative facial nerve monitoring [56–59]. Conversely, some authors believe that monitoring serves only to speed up the dissection thereby reducing the operating time substantially [5].

Various monitoring techniques have become routine in dealing with complex cranial base lesions [57]. Direct electrical stimulation, continuous electromyography (EMG), and facial motor evoked potential (fMEP) are currently useful adjuncts to CPA surgery [57–60].

Direct electrical stimulation is performed by using a bipolar stimulation probe and the responses are recorded in the subdermally placed needle electrodes into the ipsilateral orbicularis oculi and orbicularis oris. The stimuli have the purpose to directly identify the facial nerve fibers over the tumor surface [58]. Hence, dissection can be performed without risking the flattened nerve bundle [58]. Besides, the stimuli may be intermittently applied assuring the surgeon that the nerve is preserved distal to stimulation [58]. Compound muscle action potentials comparing the response from stimulation at the brainstem to the response at the internal auditory canal may be predictive of functional integrity of the facial nerve [57]. Minimal stimulus intensity of 0.05 mA or less together with a response amplitude of 240 μV or greater is able to predict House-Brackmann grades I and II outcome in 85 % of patients at 1-year follow-up [59]. Disadvantages of the method include its intermittent use and difficulty to locate the facial nerve at the brainstem in large tumors [57, 58].

Continuous EMG is better than direct electrical stimulation alone, improving the rate of facial nerve preservation during cranial base surgery [57, 58]. The spontaneous facial activity registered by EMG shows five typical waveforms that could be classified into spikes, bursts, and A, B, and C trains [58]. A-trains are sustained periodic EMG activities presenting a sinusoidal pattern that produce a high-frequency sound from the loudspeaker [58]. They occur following tumor dissection in the vicinities of the facial nerve, as well as dissection nearby the brainstem and intrameatal decompression [58]. A-trains appearance has a sensitivity of 86 % and a specificity of 89 % in demonstrating postoperative facial function deterioration [58]. In that study, however, a correlation between the number of A-trains and postoperative facial function could not be demonstrated [58]. Thereafter, Prell et al. [61] showed that a quantitative EMG parameter, namely, the train time, is a reliable indicator of postoperative facial palsy. Interestingly, 2 time thresholds are described [61]. For patients with normal preoperative function, a train time of <0.5 s is strongly correlated with a good

postoperative outcome, whereas the 10-s time threshold is associated with facial deterioration in patients with normal preoperative function and those affected by preoperative facial palsy [61].

It is worth mentioning that both studies provided an offline analysis of the intraoperative findings so that waveforms are evaluated retrospectively after the end of the surgical procedure. Recently, the study was continued by developing software capable of recognizing and quantifying A-trains in real time [62]. The software shows train time activity continuously through a monitor inlet in the surgeon’s view depicted as a traffic light display (green, < 0.125 s; yellow, 0.125–2.5 s; red, 2.5–10 s; and black, > 10 s) [62]. Such threshold levels were compared with those obtained in the offline analysis, rendering lower amounts of time by a factor of 4, in which 0.5 s of offline analysis corresponded to 0.125 s of real-time analysis. For patients with normal preoperative facial function, train time activity lower than 0.125 s was strongly correlated with good postoperative outcome, whereas 2.5-s and 10-s thresholds mostly indicated good postoperative function and marked dysfunction, respectively [62]. Conversely, for patients with preoperative facial dysfunction, only the 2.5-s threshold was defined indicating favorable or unfavorable facial function. The surgical strategy was modulated according to intraoperative findings in a way that subtotal or near-total resection was considered when the red and black thresholds were achieved. This situation occurred in 4 of 30 studied patients, whereas in the other 4 patients, threshold levels were exceeded, but the surgical procedure was continued because of the younger age [62]. That study is unique due to first-time demonstration of real-time FN monitoring with interpretation independent from the electrophysiological team. It may contribute in the near future to increasing reliability of facial function prediction during surgical procedures. Nonetheless, a major limitation is that special software is necessary for either offline or real-time analysis.

As aforementioned, both methods have limitations which have stimulated the development of alternative techniques for facial nerve monitoring. Facial motor evoked potential (FMEP), obtained by transcranial electrical stimulation (TES), has recently been introduced and may be considered the most promising method for facial nerve monitoring [57, 63–66] because it surpasses most of the disadvantages of standard techniques, namely, direct electrical stimulation and free-running EMG [65]. FMEP interpretation is independent from the surgeon’s ability to locate the proximal facial nerve at the brainstem and can facilitate recognition of waveform recordings obtained by free-running EMG [63, 64]. FMEP can be performed throughout the procedure at the demand of the surgeon, as well as following neurotonic discharges recorded by EMG or after extensive dissection on the facial nerve [57].

FMEP amplitude ratio reduction of 50 % at the end of the surgery has been identified as a good predictor for postoperative facial nerve outcome following CPA and skull-base surgeries [57, 63–65]. However, by calculating final-to-baseline amplitude ratios, only initial and final FMEP amplitude values are considered for function prediction. We hypothesized that intraoperative variation of FMEP amplitude could correlate with postoperative facial function, even if final-to-baseline amplitude ratios remain above the 50 % cutoff level. With this assumption, we have demonstrated the correlation of intraoperative FMEP deterioration with immediate and late postoperative outcome justifying changes in surgical strategy in order to avoid definitive facial nerve injury. Therefore, event-to-baseline FMEP monitoring is a better predictor of facial function than final-to-baseline isolated measurements [67].

In addition, amplitude has been the only quantitative criterion used to date for interpretation of FMEP [67]. Thus, we also investigated the intraoperative changes of the FMEP waveform morphology and their correlation to postoperative facial function [67]. We found that waveform complexity seems to represent an additional, independent quantitative parameter that can be used for FMEP monitoring, providing excellent results for prediction of postoperative facial function [67].

Just as for cochlear nerve monitoring, the surgical steps are modified or even interrupted if facial nerve changes happen that are known to be significant in postoperative deterioration of facial function. Table 23.4 summarizes prediction criteria for facial nerve monitoring which are related to a satisfactory functional outcome.