The development and global implementation of cochlear implants (CIs) over the past 30+ years has dramatically improved the quality of life for patients with moderate to profound sensorineural hearing loss (1). Unfortunately, a number of patients suffer loss of residual hearing following CI (2–4). While the cause of residual hearing loss is multifactorial (5,6), a critical contributing factor is trauma sustained to the cochlea during insertion of the electrode array. Atraumatic CI electrode array insertion is associated with improved outcomes even in patients with severe-profound hearing loss (7,8).
Preservation of inner ear structures is paramount to preserving residual hearing and improving outcomes with cochlear implantation. Methods to improve structure preservation and residual hearing have included refinements to electrode array design (reduced diameter, smooth tip, perimodiolar versus lateral wall) (9), surgical approach (cochleostomy versus round window [RW] insertion), use of lubricants or drugs in the cochlea (e.g., intraoperative corticosteroids) (10,11), and a general emphasis on the use of atraumatic or “soft” surgical techniques during the electrode array insertion. Despite increasing surgeon awareness of the importance of soft surgical techniques, there remains a high variability of hearing preservation outcomes across surgeon experience, patient demographics, implant centers, and electrode types (4,12–16). In addition, multiple studies have indicated ∼50% of hearing preservation implant recipients experience additional loss of their natural hearing (>10 dB) in the months and years following surgery (13,17–19).
To address the issue of intracochlear insertion trauma, a robotics-assisted surgical tool was developed to aid the surgeon in CI electrode array insertion. The open architecture system was designed to be compatible with current surgical approach and electrodes from multiple CI manufacturers to assist the surgeon in more precisely and consistently inserting the electrode array. The objective of this study was to compare insertion forces and cochlear trauma between robotics-assisted insertions and manual insertion techniques. Here we evaluate the differences in insertion forces in benchtop and ex vivo human cadaveric models as well as compare intracochlear insertion trauma in human cadaveric cochleae using a novel, non-destructive imaging technique.
Benchtop Insertion Forces
CI electrode array insertions were performed through the RW either by hand with jeweler forceps or utilizing a robotics-assisted insertion system (iotaMotion Inc, Iowa City, IA). Test insertions were conducted in both a 3D printed synthetic cochlea (20) (Stratasys, Polyjet HD, Los Angeles, CA) and embalmed, explanted human cadaveric cochleae (n = 3). The cochleae were secured to single-axis force transducer (Futek LRF400, Irvine, CA) orthogonal to long axis of the cochlea and submerged in buffered saline during testing.
Four different lateral wall electrode array types (Cochlear, Advanced Bionics, MedEl, Oticon) were inserted by a single operator either manually with forceps or with the robotics-assisted insertion system at 0.1, 0.5, and 1.0 mm/s, repeated in triplicate and utilizing a new electrode for each insertion. For manual insertion, the insertion rate was targeted based on total insertion time with a constant insertion motion attempted. Care was taken to insert each electrode at the same trajectory into the cochleae along the floor of the scala tympani while maintaining at least 2 points of hand/arm fulcrum stabilization.
Insertions were analyzed with respect to the maximum peak force (mN) and force variation (mN/s). Maximum peak force was the single highest force observed during that insertion. Force variation was calculated as previously described by Nguyen et al. (21). Comparisons between manual and robotics-assisted insertions for each speed, electrode array, and cochlea model were determined using a t test (p < 0.05) and analysis within each of the insertion speeds performed by analysis of variance with post hoc analysis via Tukey's HSD (p < 0.05).
Human Cadaveric Cochlea Insertion Trauma: Manual Versus Robotics-Assisted
The robotics-assisted insertion system used in this study was designed to be used within the current surgical framework of standard cochlear implantation via facial recess approach while maintaining standard operating microscope view of the surgical field and cochlea. The test system is a powered, surgeon-controlled device that consists of a compact insertion drive unit operated via surgeon foot pedal. At the time of electrode array insertion, the surgeon temporarily fixes the drive unit to the skull at edge of mastoidectomy with bone screws (Fig. 1A). The unit drive head is then coupled to the CI electrode array excess lead via spring-loaded, split clamping mechanism. The drive head and split guide sheath tips are then positioned and aligned by manual manipulation of the semi-rigid extension arm to achieve the surgeon-desired electrode insertion trajectory (Fig. 1B). While maintaining direct microscopic view of the surgical site and instrument access as needed, the surgeon actively advances the electrode array into the cochlea via footpedal at controlled insertion rates and distances with resolution of 0.1 mm. The electrode array insertion was continuously visualized under the operating microscope (Zeiss OPMI 111) for any undesired movements and standard forceps or claw instrument was used to make any fine insertion trajectory adjustments as desired by the surgeon.
Fresh frozen human temporal bones (n = 24) were thawed at room temperature (RT) for 6 to 12 hours then submerged in RT water for approximately 1 hour before use. Standard mastoidectomy and facial recess approach was performed under operating microscope (Zeiss OPMI 111) with the RW niche removed until a complete view of RW membrane was obtained. Electrode arrays were inserted through the RW either by hand (n = 12) with jeweler forceps or utilizing the robotics-assisted systems as describe above (n = 12) using a fresh temporal bone cochlea for each insertion, repeated in triplicate for each the four electrode array types. Varying surgical skill levels from single institution (University of Iowa; neurotology fellow, senior level resident, and experienced CI surgeon) were randomized for each manual electrode array insertion and surgeons were instructed to insert the array “atraumatically.” A single surgeon performed the robotics-assisted insertions for each electrode type. If significant resistance was felt during manual insertion or buckling of the proximal extracochlear electrode array was observed during robotics-assisted insertion before complete electrode array insertion, the surgeon stopped the insertion without attempting to insert further past resistance.
Following insertions, excess fluid was gently suctioned from cochlea and electrode arrays were fixed in place with methacrylate glue just external to RW niche. The otic capsule was carefully explanted and samples were imaged with electrode array in situ via high resolution x-ray microscopy (Zeiss Xradia 520 Versa, 8–12 μm pixel size, Carl Zeiss Microscopy, Jena, Germany). The electrode array was then removed while maintaining sample position and the cochlea re-imaged. Using ORS visual software suite (Montreal, Canada) the respective images both with and without electrode array were overlaid to create a 3D image reconstruction with intact cochlea tissue details and implant in situ position without associated implant metal artifact. Once alignment registration was confirmed the images were analyzed using available 2D and 3D suite tools.
X-ray Microscopy Trauma Assessment and Scoring
X-ray microscopy (XRM) was used to directly assess the insertion trauma associated with each robotics-assisted and manual insertion according to a custom scale specific to this XRM imaging modality and clinically relevant severity (Table 1). Each cochlea (n = 24) was scored according to the most severe event level observed (worst case observation). Statistical analysis between manual and robotics-assisted insertion trauma scoring was performed across insertion trauma events using a Likelihood Ratio χ2 test.
Maximum Insertion Force
The maximum forces across all aggregated electrode arrays for manual and robotics-assisted insertions are shown in (Fig. 2). When electrode arrays were inserted into either synthetic (Fig. 2A) or cadaveric (Fig. 2B) cochleae, the robotics-assisted insertions had significantly lower maximum insertion forces at all test speeds compared with manual insertions in both synthetic and cadaveric cochlea. Within the different insertion speeds, 0.1 mm/s had significantly higher maximum insertion forces for both manual and robotics-assisted insertions when compared with either 0.5 or 1.0 mm/s. For individual electrode array maximum insertion forces, see Supplemental Figure 1, http://links.lww.com/MAO/A942.
The force variation across all aggregated electrode array insertions using robotics-assisted and manual insertions is shown in Fig. 3. When electrodes were inserted into either synthetic (Fig. 3A) or cadaveric cochleae (Fig. 3B), use of the robotics-assistance significantly reduced the force variation at every speed tested compared with manual insertions. Additionally, there was a statistically significant difference between each speed with robotic-assisted insertion in cadaveric cochlea. Robotics-assisted insertions in cadaveric test specimens at 0.1 mm/s exhibited the lowest variation compared with 1.0 mm/s which had the highest force variation. For individual electrode array insertion force variation, see Supplemental Figure 2, http://links.lww.com/MAO/A943 (Fig. 4).
Insertion Trauma Scoring
Representative x-ray microscopy (XRM) images of identifiable trauma events are shown in Figure 4 with the least to most severe event shown based on Table 1. For manual insertions, there was a trend toward more severe insertion trauma events with more insertions having isolated scala vestibuli (SV) translocations (8% versus 25%) and OSL fractures (0% versus 17%) as the most severe trauma event. Manual insertions had significantly higher rate of insertions with combined OSL fracture and electrode translocation compared with robotics-assisted insertions with no level 5 events (0% versus 33%, p = 0.012). Robotics-assisted insertions exhibited a lower incidence of detectable trauma events compared with manual insertions (42% versus 17%). There was a higher percentage of insertions with less severe trauma events of basilar membrane (BM) elevation (33% versus 8%) and scala media (SM) translocations (17% versus 0%) for robotics-assisted insertion compared with manual electrode array insertions as shown in Figure 5.
There was no significant difference in electrode array insertions angles (depth) between manual and with robotics-assistance, 359 ± 113 degrees compared with 321 ± 84 degrees, respectively. All isolated OSL fractures occurred at an overall average insertion angle of 25 ± 10 degrees while basilar membrane elevation and translocation events were identified at 159 ± 61 degrees and 161 ± 34 degrees, respectively. There was no significant difference in the trauma event locations between manual and robotics-assisted insertions.
Effect of Manual Versus Robotics-Assisted Insertion
Use of the robotics-assisted system in this study significantly reduced the observed CI electrode array maximum insertional forces, force variations, and intracochlear trauma score when compared with manual CI electrode array insertions. Electrode arrays inserted with robotics-assistance had a significant reduction in insertional force variation for all electrode arrays, speeds, and cochlea models compared with manual insertions. Force variation is a relative measure of insertion profile smoothness (21) and can be described as a relative characterization of the number and magnitude of force peaks seen throughout a particular insertion. Each peak represents a significant rise in force with the potential to damage intracochlear structures. Intermittent insertion starting and stopping associated with manual insertions resulted in significantly more force peaks and potential for damage from higher maximum force spikes. The decreased force variation with robotics-assistance is one possible mechanism for the reduced intracochlear trauma event severity and scores seen in this study.
Another potential benefit of reduced force variation may arise from the changes in intracochlear pressure that occur during the electrode insertion. Although pressure changes were not evaluated in the current study, studies indicate that increased intracochlear pressure may cause intracochlear damage and is positively correlated with insertion speed (22). A recent study by Banakis Hartl et al. (23) showed that potentially harmful transient pressure spikes could be reduced through use of a micromechanical electrode insertion device at insertion speeds similar to those used in the current study. If pressure spikes are the result of rapid motion of the electrode into the cochlea without sufficient time to allow for pressure equilibration, reducing these pressure spikes may be possible by reducing the incidence of rapid changes in the speed of the electrode array advancement. These acceleration events can be determined indirectly by the force variation metric in the current study, and are significantly reduced by use of a robotics-assisted electrode array insertion.
The speed of electrode array insertion has been shown previously to positively correlate with increased insertional forces (24–27). Insertion speeds of 0.1, 0.5 and 1.0 mm/s in this study were chosen to represent the low end and average of typical reported manual insertion speeds (27) as well as constant, robotics-assisted speeds which are below human limits (26). In this study, insertion speed did not appear to have a substantial effect on the observed maximum forces in either synthetic or cadaveric test specimens which is likely due to testing such low speed range. However, at the slowest insertion speed of 0.1 mm/s in synthetic cochleae only, there were significantly higher maximum insertion forces for both manual and robotic insertions. The reason for this reverse correlation finding at the slowest insertion speed in synthetic cochlear is unclear and additional studies are needed to elucidate any potential difference in electrode insertion dynamics at ultra-low insertion speeds less than 0.1 mm/s. Importantly, this observation may be an artifact of material interactions in synthetic cochleae as it was not observed in cadaveric cochleae.
When compared with manual insertions, the use of the robotics-assisted system influenced the observed maximum insertion force on the cochlea, a common metric used to evaluate electrode insertions performed in a laboratory setting (24,28,29). As the insertion of an electrode array progresses, friction encountered necessitates a higher force to push the electrode array deeper into the cochlea. In the current study, when averaged across all electrodes, robotics-assisted insertions had significantly lower maximum forces at every speed. A constant insertion rate may limit static friction effect during the insertion, thereby diminishing the force necessary to advance the electrode array deeper into the cochlea. Robotics-assisted insertion can also prevent the surgeon from applying excessive force when attempting to overcome the increasing frictional forces during electrode array advancement. Such efforts to overcome resistive frictional forces appear to result in large force spikes and increased potential for trauma including translocation of the array out of scala tympani during manual insertion.
CI insertion trauma has traditionally been assessed by histological sectioning followed by analysis of individual slides or section images. The CI electrode array must be removed before decalcification, embedding, and microtome sectioning or if the electrode array is kept in place, epoxy resin embedding is required followed by destructive grinding for image sectioning. With the advancement of imaging technology, we have demonstrated a novel means for non-destructive analysis of electrode insertion trauma with electrode array maintained in situ using a commercially available high-resolution 3D x-ray microscopy (XRM) system. With image resolution ranging from 6 to 12 μm, there was sufficient soft and hard tissue resolution and contrast to adequately distinguish intracochlear tissues and structures on a macroscopic level including basilar membrane, Reisner's membrane, spiral ligament, and osseous spiral lamina. The XRM technique is feasible for rapid, sample analysis with an average scan time of approximately 1 hour per sample, and no previous sample preparation or processing required.
Two scans were used to diminish the effect of electrode metal artifact on soft tissue images; one with the electrode array in place and a second after removal. This allowed window-leveling out of the electrode artifact signal followed by overlay and co-registration onto the second scan of native cochlea soft tissue while maintaining intracochlear electrode position from initial scan. The current XRM technique allows for microstructure resolution with electrode position maintained in situ but has the downside of requiring electrode array removal. It is possible the action of removing the electrode results in trauma during removal. However, to mitigate and monitor for this, any trauma event identified in overlaid scans was also reviewed in the first scan to verify presence, if possible, before electrode withdrawal. While significant metal artifacts are present in the initial scans, sufficient detail remained to allow assessment and conclusion that no identifiable trauma events seen in the co-registered images were present that were not also seen in the initial scan. Future strategies may lead to improved metal artifact correction, yet these techniques are based on image correction algorithms rather than structural x-ray attenuation patterns. Additionally, dual energy imaging has potential to mitigate metal artifacts but in our experience this technique required significantly longer scan times with marginal improvement in image quality.
In this study the fresh frozen cadaveric tissue was not fixed or preserved either before or after electrode removal to most closely simulate live human insertions. Without tissue fixation it is possible that the intracochlear structures could shift position after electrode removal. This would have the potential effect of skewing results toward more severe trauma event scoring of basilar membrane elevation or scala translocation. For example, upon initial insertion the electrode may have elevated, but not ruptured the BM. The first scan provides the electrode intracochlear position. Upon removal of the electrode, the elevated BM could drop to native position and the sample scan is repeated. When the images are overlaid, based on electrode position (first scan) and location of BM and SM, the electrode position may appear to be within the SM rather than in the ST with elevated BM. This structure shift phenomenon associated with our current XRM imaging technique may be mitigated in future studies by chemically fixing the cochlea in formalin following the insertion. The resulting increase in intracochlear tissue mechanical strength and stability would help prevent a shift in the intracochlear structures’ position following electrode removal.
Insertion trauma is a broad categorization of a number of unique events that affect various aspects of the inner ear. Excessive forces exerted by the electrode array onto lateral wall tissues can disrupt or tear the BM or translocate the electrode into the SM or SV (30–32). Additionally, elevation of the BM by the electrode array may interfere with the natural vibrational mechanics as sounds travels towards the apex, directly compromising residual hearing (33). Lastly, fractures of the osseous spiral lamina (OSL) can sever the peripheral fibers of spiral ganglion cells and lead to ganglion cell degeneration (34,35). Thus, confining electrode array placement to the scala tympani (ST) and preservation of inner structures is paramount to preserving residual hearing and improved outcomes with cochlear implantation.
For trauma assessment, we adapted a previous Roland and Wright (33) electrode insertion trauma scale to the high resolution XRM modality. The expanded scale (Table 1) accounts for estimated clinical significance of the spectrum of trauma events commonly identified on XRM. Specifically, we assigned lower values to events which did not violate inner ear structures (BM elevation) versus those which did (SM and SV translocation, OSL fracture). This is of relevance to this study, as we hypothesize the former events involving translocation or fracture may require a higher insertion force in addition to have more dramatic physiologic consequences in the setting of traditional and hearing-preservation cochlear implantation. However, further studies may be needed in the future enhance the validity and clinical relevance of such a scale.
In this study a significant number of OSL fractures with and without electrode translocation were identified. These level 5 events were characterized to account for the two scenarios for translocation in clinical setting—either through the basilar membrane or via fracturing through the OSL. We think fracture of the OSL with translocation is a more severe event. However, it is unclear if the significant OSL fractures identified represent clinically present and relevant trauma events or are the result of weakened cadaveric tissue. We suspect these microfractures likely occur in clinical practice yet there is no current live imaging modality which has sufficient resolution to detect OSL microfractures in vivo. Given the high resolution of the XRM imaging technique and ability to review and analyze multiple section orientations, these microfracture may not be identified with traditional histology of postmortum studies due to 3D spatial limitations of traditional histologic technique and the timing between injury and postmortum examination. If an acute OSL microfracture occurs in a chronic in vivo setting, an inflammatory and osteogenic process may ensue to repair the damage. Unless this trauma event is analyzed acutely, histological evidence of the initial fracture may not be detected except for non-specific remodeled bone or tissue. It is possible that XRM detects previously unidentified microfractures that are known to be a significant mechanism for intracochlear inflammation and trauma (34,35).
The current study supports the use of a robotics-assisted systems for CI electrode array insertions with potential to reduce insertion trauma to the cochlea. This study was primarily designed to characterize and evaluate benchtop and ex vivo outcomes of electrode array insertion with robotics-assisted and manual insertions. Future, studies will seek to evaluate and compare surgical techniques for reduced trauma and improved outcomes in an acute and chronic large animal model. Additionally, the use of XRM is an effective and non-destructive method for evaluating intracochlear trauma following CI electrode array insertion and could be used in future in vivo experiments to more accurately quantify the level of trauma and correlate these observations with functional hearing outcomes.
As the population of nearly every developed country continues to age at a rapid pace, the use of cochlear implants to treat hearing loss will continue to rise. Methods to precisely control and standardize CI electrode array insertions across a spectrum of surgical experiences, such as robotics-assisted devices, holds the potential to improve outcomes for patients undergoing cochlear implantation, as well as increase the candidacy range for CIs by increasing the probability of residual hearing preservation.
Michael Acevedo and Susan Walsh at the Small Animal Imaging Core, University of Iowa College of Medicine, Department of Radiology and Nuclear Medicine for technical assistance on Xradia system (NIH 1S10OD018503-01). Adam Hahn and Parker Reineke for technical design assistance of insertion system. Tom Roland, Craig Buchman, and Bruce Gantz for surgical guidance. Cochlear, Advanced Bionics, MedEl, and Oticon for supplying test electrodes.
1. Goman AM, Lin FR. Prevalence of hearing loss by severity in the United States. Am J Public Health
2. Gantz BJ, Dunn CC, Oleson J, Hansen MR. Acoustic plus electric speech processing: long-term results. Laryngoscope
3. Scheperle RA, Tejani VD, Omtvedt JK, et al. Delayed changes in auditory status in cochlear implant
users with preserved acoustic hearing. Hear Res
4. Van Abel KM, Dunn CC, Sladen DP, et al. Hearing preservation among patients undergoing cochlear implantation. Otol Neurotol
5. Kopelovich JC, Robinson BK, Soken H, et al. Acoustic hearing after murine cochlear implantation: effects of trauma and implant type. Ann Otol Rhinol Laryngol
6. Kopelovich JC, Reiss LAJ, Oleson JJ, Lundt ES, Gantz BJ, Hansen MR. Risk factors for loss of ipsilateral residual hearing after hybrid cochlear implantation. Otol Neurotol
7. Carlson ML, Driscoll CLW, Gifford RH, et al. Implications of minimizing trauma during conventional cochlear implantation. Otol Neurotol
8. Cosetti MK, Friedmann DR, Zhu BZ, et al. The effects of residual hearing in traditional cochlear implant
candidates after implantation with a conventional electrode. Otol Neurotol
9. Jolly C, Garnham C, Mirzadeh H, et al. Electrode features for hearing preservation and drug delivery strategies. Adv Otorhinolaryngol
10. Eshraghi AA, Adil E, He J, et al. Local dexamethasone therapy conserves hearing in an animal model of electrode insertion trauma
-induced hearing loss. Otol Neurotol
11. Lee J, Ismail H, Lee JH, et al. Effect of both local and systemically administered dexamethasone on long-term hearing and tissue response in a guinea pig model of cochlear implantation. Audiol Neurotol
12. Gstoettner WK, Van de Heyning P, Fitzgerald O’Connor A, et al. Electric acoustic stimulation of the auditory system: results of a multi-centre investigation. Acta Otolaryngol
13. Gantz BJ, Dunn C, Oleson J, Hansen M, Parkinson A, Turner C. Multicenter clinical trial of the Nucleus Hybrid S8 cochlear implant
: final outcomes. Laryngoscope
14. Eshraghi AA, Lang DM, Roell J, et al. Mechanisms of programmed cell death signaling in hair cells and support cells post-electrode insertion trauma
. Acta Otolaryngol
15. Garcia-Ibanez L, Ramos Macias A, Morera C, et al. An evaluation of the preservation of residual hearing with the Nucleus®
Electrode. Acta Otolaryngol
16. Brown KD, Melton MF, Shonfield H, Kraskin M, Wolf J. Preserved low-frequency hearing following 20-mm cochlear implantation. Otol Neurotol
17. Gantz BJ, Hansen MR, Turner CW, Oleson JJ, Reiss LA, Parkinson AJ. Hybrid 10 clinical trial: preliminary results. Audiol Neurootol
2009; 14: (suppl): 32–38.
18. Roland JT, Gantz BJ, Waltzman SB, Parkinson AJ. United States multicenter clinical trial of the cochlear nucleus hybrid implant system. Laryngoscope
19. Lenarz T, James C, Cuda D, et al. European multi-centre study of the Nucleus Hybrid L24 cochlear implant
. Int J Audiol
20. Leon L, Cavilla MS, Doran MB, Warren FM, Abbott JJ. Scala-tympani phantom with cochleostomy and round-window openings for cochlear-implant insertion experiments. J Med Device
21. Nguyen Y, Kazmitcheff G, De Seta D, Miroir M, Ferrary E, Sterkers O. Definition of metrics to evaluate cochlear array insertion forces
performed with forceps, insertion tool, or motorized tool in temporal bone specimens. Biomed Res Int
22. Todt I, Mittmann P, Ernst A. Intracochlear fluid pressure changes related to the insertional speed of a CI electrode. Biomed Res Int
23. Banakis Hartl RM, Kaufmann C, Hansen MR, Tollin DJ. Intracochlear pressure transients during cochlear implant
electrode insertion: effect of micro-mechanical control on limiting pressure trauma. Otol Neurotol
24. Majdani O, Schurzig D, Hussong A, et al. Force measurement of insertion of cochlear implant
electrode arrays in-vitro: comparison of surgeon to automated insertion tool. Acta Otolaryngol
25. Miroir M, Nguyen Y, Kazmitcheff G, Ferrary E, Sterkers O, Grayeli AB. Friction force measurement during cochlear implant
insertion: application to a force-controlled insertion tool design. Otol Neurotol
26. Kesler K, Dillon NP, Fichera L, Labadie RF. Human kinematics of cochlear implant
surgery: an investigation of insertion micro-motions and speed limitations. Otolaryngol Neck Surg
27. Rajan GP, Kontorinis G, Kuthubutheen J. The effects of insertion speed on inner ear function during cochlear implantation: a comparison study. Audiol Neurotol
28. Kontorinis G, Lenarz T, Stöver T, Paasche G. Impact of the insertion speed of cochlear implant
electrodes on the insertion forces
. Otol Neurotol
29. Mirsalehi M, Rau TS, Harbach L, et al. Insertion forces
and intracochlear trauma in temporal bone specimens implanted with a straight atraumatic electrode array
. Eur Arch Otorhinolaryngol
30. Duvall A, Rhodes V. Ultrastructure of the organ of Corti following intermixing of cochlear fluids. Ann Otol Rhinol Laryngol
31. Shaddock LC, Wright CG, Hamernik RP. A morphometric study of microvascular pathology following experimental rupture of Reissner's membrane. Hear Res
32. Axelsson A. The vascular anatomy of the cochlea in the guinea pig and in man. Acta Otolaryngol
1968; 243: (suppl): 1–134.
33. Roland PS, Wright CG. Surgical aspects of cochlear implantation: mechanisms of insertional trauma. Adv Otorhinolaryngol
34. Leake PA, Hradek GT, Snyder RL. Chronic electrical stimulation by a cochlear implant
promotes survival of spiral ganglion neurons after neonatal deafness. J Comp Neurol
35. Leake PA, Rebscher SJ. Zeng F-G, Popper AN, Fay RR. Anatomical Considerations and Long-Term Effects of Electrical Stimulation BT - Cochlear Implants: Auditory Prostheses and Electric Hearing. New York, NY:Springer New York; 2004. 101–148.