*2.5. Statistical Analyses*

The proportion of EMG responses with ST1 and ST2 were compared using the Mc-Nemar's test. Spearman's correlation test was used to investigate associations between intraoperative EMG responses and cochlea-nerve distances estimation of the subjects. The correlation between the electrode-nerve distances estimation of the electrode E15 (placed in the range closest to the facial nerve in most subjects, from 250 to 290 degrees) and T-FNS were also investigated. The results were expressed in correlation coefficient (ρ) and *p*-value. A significance level of 5% was adopted.

#### **3. Results**

Table 3 shows the proportion of intraoperative EMG responses recorded on each tested electrode using the stimulation patterns ST1 and ST2. ST1 stimulation leaded to intraoperative EMG responses in at least one electrode in 9 of 10 ears while the ST2 stimulation induced EMG responses only in the most basal electrode (E1) in 1 of 10 ears (#S4). Subject #S3 (EA6) showed absent EMG recordings with both stimulation patterns ST1 and ST2 (with orbicularis oculi and oris muscles input channels). Overall, the paired analyses indicated that the use of ST2 was significantly associated to lower incidence of intraoperative EMG responses recorded on all tested electrodes compared to ST1.

**Table 3.** Proportion of intraoperative EMG responses recorded in each tested electrode using the stimulation patterns ST1 and ST2.


ST1: stimulation pattern 1 (monopolar biphasic stimulation); ST2: stimulation pattern 2 (multi-mode monophasic stimulation); N: number of EMG responses in each implanted ear. \* Significant difference (McNemar's test, 5% of significance level).

Figure 3 shows individual T-FNS (A) and EMG amplitudes (B), when recorded, using CI stimulation patterns ST1 and ST2. Two subjects showed EMG responses only in the electrode E1 (#S5 with ST1 and #S4 with ST2), and their absence in the remaining electrodes. Higher stimulation levels were required to elicit FNS (T-FNS) with ST2 (37 nC/phase) compared to ST1 (13 nC/phase) in the subject #S4 (electrode E1). Furthermore, peakto-peak EMG amplitudes were smaller using ST2 (15 μV) compared to ST1 (27 μV) in this electrode.

**Figure 3.** Individual T-FNS (**A**) and EMG amplitudes (**B**), when recorded, using CI stimulation patterns ST1 and ST2. Blank: no EMG responses. EA1–EA10: implanted ears 1–10. \*Asterisks: comparison between values recorded with ST1 and ST2, in the same electrode (#Subject 4, EA7).

The intra-cochlear electrodes placement and cochlea-nerve distances estimation are provided in the Table 4. Four subjects (six ears) showed at least one extra-cochlear electrode placement (basal electrodes), including the subject #4, who showed EMG responses only in the most basal electrode with ST2 (Figure 4). The cochlea-nerve distances ranged from 0.20 to 1.00 mm (mean = 0.42 ± 0.26 mm). One subject (#S3, EA6) showed the longest cochlea-nerve distance and absent intraoperative EMG recordings with both the ST1 and ST2, but no correlation was observed between cochlear-nerve distances and EMG responses (Table 5). E15 and E16 were the Evo® electrodes placed closest to the labyrinthine segment of the facial nerve, being recorded from 250 to 290 degrees insertion depth angle in most subjects. The electrode-nerve distance from the electrode E15 showed a weak correlation with the T-FNS recorded on this electrode (*ρ* = 0.42; *p*-value = 0.30).

**Table 4.** Intra-cochlear electrodes placement, cochlea-nerve and electrode-nerve distances estimation of the subjects.


S1–S7: subjects 1–7; EA1–EA10: implanted ears 1–10; L: left; R: right; cochlea-nerve distance: closest distance between basal turn of the cochlea and the labyrinthine segment of the facial nerve; mm: millimeters; Electrode closest to the facial nerve: Evo® electrode with closest electrode-nerve distance values; FN: facial nerve.

**Figure 4.** Example of electrodes placement estimation performed by Nautilus for the subject #S4 (EA7), who showed FNS responses with ST1 and ST2, in the electrode E1. Electrode 1 (basal) is extra-cochlear. \* Blue circle and asterisks correspond to the labyrinthine segment of the facial nerve and Evo® electrodes with closest electrode-nerve distances, respectively.

**Table 5.** Relationship between cochlear-nerve distances and EMG responses of the subjects.


nC: Nanocoulomb; μV: microvolt; T-FNS: facial nerve stimulation thresholds; EMG Amplitude: peak-to-peak electromyographic amplitudes. Spearman's correlation test, at a significant level of 5%.

#### **4. Discussion**

The purpose of this study was to investigate the effects of the CI electrical stimulation pattern on FNS reduction in young children. We recorded intraoperative EMG action potentials in children implanted up to 6 years old, using two different CI electrical stimulation patterns: monopolar biphasic stimulation (ST1) and multi-mode monophasic stimulation with capacitive discharge (ST2). Presence of EMG responses, T-FNS and EMG amplitudes were compared between them. Multi-mode monophasic stimulation with capacitive discharge significantly reduced the presence of EMG responses compared to monopolar biphasic stimulation using equal stimulation levels in young children with CI.

To the best of our knowledge, this is the first study comparing intraoperative EMG action potentials between ST1 and ST2 electrical stimulation patterns in CI children recipients and providing electrophysiological evidence of ST2 on FNS reduction in this population. Similar results were recently reported by Eitutis et al. [7] in three adult patients re-implanted with the Neuro Zti device due to severe FNS. Intraoperative FNS were recorded in all the three subjects with ST1, while no FNS was observed with ST2. In our study, FNS was recorded only in the most basal electrode of the subject #S4 with ST2, and the 3D image reconstruction revealed that this electrode was extra-cochlear (Table 4, Figure 4). It is known that extra-cochlear electrodes increase the risk of FNS in CI recipients [15,16], since the total amount of current spread from electrodes which lie outside the cochlea to the periphery structures, including the facial nerve. Thus, considering only the electrodes placed inside the cochlea and, therefore, excluding the extracochlear electrodes from the analysis, the FNS was not recorded with ST2 in any of the subjects included in this study. Our results reinforce those found by Eitutis et al. [7] in adults and suggest that the stimulation pattern ST2 seems to have comparable effects on FNS reduction in the first CI implantation of young children.

Higher stimulation levels were required to elicit FNS with ST2, with smaller EMG amplitudes, compared to ST1. This is also the first time that EMG input-output functions [9] (i.e., charge level required to elicit FNS versus EMG amplitude) could be compared between ST1 and ST2. The CI stimulation pattern ST2 corresponds to the combination of both, the multi-mode grounding stimulation and anodic monophasic pulse shape with capacitive discharge. The relative contribution of each of these features in FNS reducing in CI recipients has been unclear, since the software Genie Medical CI (Oticon Medical, Denmark) does not allow their dissociation for stimulation. In the multi-mode grounding stimulation, a greater amount of electrical current is maintained inside the cochlea since it flows from a stimulating intra-cochlear electrode to the remaining non-stimulating intra-cochlear electrodes. Considering that the electrode E1 (#S4) was extra-cochlear, the comparison of the EMG input-output functions between ST1 and ST2 seems to have been mainly influenced by the pulse shape, since the total amount of the electrical current was spread outside the cochlea, similarly to the monopolar stimulation mode pathway.

The distance between the basal turn of the cochlea and the labyrinthine segment of the facial nerve in the subjects ranged from 0.20 to 1.00 mm. These results are in accordance with previous studies on human temporal bones based on histological image measurements or macroscopical/microscopical analysis [17,18] and CI adult recipients based on preoperative CT scans analysis (axial and coronal orientation plan) [4]. Hatch et al. [4] also investigated the effects of the cochlea-nerve distance on FNS in 49 ears of adult CI recipients and found lower cochlea-nerve distances in subjects with FNS compared to a control group (with no FNS). They suggested that cochlea-nerve distances longest than 0.6 mm should decrease the risk of FNS in this population. In our study, no correlation was observed between cochlear-nerve distances and intraoperative EMG responses in young children with CI, but one subject (#S3, EA6) showed cochlea-nerve distance longer than 0.6 mm, and only this subject showed absent intraoperative EMG recordings with both the ST1 and ST2. The Evo® electrodes E15 and E16 were the closest to the labyrinthine segment of the facial nerve, and they were placed in the upper basal turn of the cochlea, from 250 to 290 degrees insertion depth angle, in most subjects (Figure 4). Seyyedi et al. [19] supposed

that electrodes placed in the upper basal turn of cochlea should be most likely to excite the facial nerve, due to their proximity to it. Our findings may confirm the closest proximity of the electrodes placed in the upper basal turn of cochlea to the labyrinthine segment of the facial nerve, but only a weak correlation was observed between the electrode-nerve distance and intraoperative FNS responses, based on our results to the electrode E15. One limitation of this analysis was the small sample size (N = 10), nevertheless, strong correlations between these factors should be detected using the Spearman's correlation test at *p* < 0.05. Even though, further investigations with a larger number of subjects are required to explore the results of intra-cochlear electrodes positioning to better understand the relative influence of this factor on FNS reduction in this population. Anyway, the use of 3D image processing techniques allowed us to accurately estimate the CI electrodes positioning and electrodenerve distances based on the real intra-cochlear electrodes' placement. This analysis was essential and provided better investigation on these aspects, considering the high variability in the CI electrodes positioning [20,21].

In this study, the EMG recordings were carried out under general anesthesia, and, as muscle relaxants could affect the EMG responses [9], they were not administrated, and thereby our findings may be compared to the clinical routine of non-anesthetized patients.

Finally, our results suggest that CI electrical stimulation pattern may affect the FNS in young children and multi-mode monophasic stimulation with capacitive discharge should far reduce FNS in young children with CIs. The adoption of this electrical stimulation pattern should be an effective option for patients with a higher risk of experiencing FNS after CI surgery, such as patients with otosclerosis, meningitis, temporal bone fractures and congenital cochlear anomalies, or those who have indication for cochlear re-implantation due to severe FNS.
