Abstract

This paper presents a platform for self-learning cochlear insertion using computer vision in a three-dimensional surrogate model. Self-learning and practice experiences often improve the confidence associated with eventual real-world trials by novice medical trainees. This helps the trainees practice electrode insertion to minimize the effect of suboptimal electrode placement such as incomplete electrode insertion, electrode kinking, and electrode tip fold-over. Although existing mastoid fitting templates improve insertion trajectories, extensive training is still required. Current methods that use cadavers, virtual training, or physical models from reconstruction images are not good enough for training purposes. The model presented here simulates the dimensions, texture, and feel of inserting the electrode into the cochlea. Currently, the temporal bone is not included, hence it is not meant for practicing drilling and other procedures to access the cochlear. The insertion process is observed in real-time using a camera and a Graphical User Interface that not only shows the video feed, but also provides depth, trajectory, and speed measurements. In a trial conducted for medical trainees, there was an overall improvement in all four metrics after they were trained on the hardware/software. There was a 14.20% improvement in insertion depth, 44.24% reduction in insertion speed, 52.90% reduction in back-outs, and a 64.89% reduction in kinks/fold-overs. The advantage of this model is that medical trainees can use it as many times as they like, as the whole setup is easy, economical, and reusable.

Introduction

Sensorineural hearing loss, or nerve-related hearing loss, involves issues with the inner ear, and can occur in one or both ears. Hearing aids can be used to correct this if there are some remaining healthy sensory hair cells left in the cochlea that can transmit sound to the brain using amplification. However, when this is not the case, surgical treatment with cochlear implants (CI) is a typical option [13]. A cochlear implant is a small electromechanical device that electrically stimulates the cochlear nerve through the cochlea [4]. Unlike hearing aids which only amplify sounds, CIs bypass damaged portions of the ear to deliver an electrical signal to the auditory nerve which the brain learns to interpret as sound [5]. The CI has two main components: an external device (the processor) and an internal device (the implant). The internal device is placed surgically, typically during outpatient surgery. It houses a removable magnet and an electrode array. The electrode array is inserted into a specific part of the cochlea called the scala tympani. In general, to insert the electrode the surgeon makes an incision behind the ear and opens the mastoid bone to access the inner ear. The surgeon drills through the temporal bone with progressively smaller drill bits to reach the round window. An incision is made in the round window membrane to thread the electrode down into the cochlea. Next, the surgeon mounts the internal piece behind the ear and beneath the skin and secures it to the skull. The surgeon then closes the incisions.

Studies have shown that the overall complication rate for CIs is 19.9%, comprising 5% of major cases and 14.9% of minor cases [6,7]. Typical complications include infection in the area around the implant, injury to the facial nerve, CI migration from its osseous bed, electrode array dislodgement from the cochlea, and suboptimal electrode placement such as an incomplete electrode insertion, kinking, and tip fold-over [810]. Usually, CI surgery is relatively safe in the hands of experienced implant surgeons [11], but they need years of experience dealing with different cases, patients, and implants. Although devices such as mastoid fitting templates have been developed to improve electrode insertion trajectory [12], extensive training is still required to gain the needed expertise to lower the complication rates for this procedure. Key to lowering the complication rate for CI surgery is training surgeons to expert level without having to risk patient outcomes [13].

There are three methods currently available for CI surgery training: use of a cadaver/temporal bone, use of a virtual reality (VR) system, and use of three-dimensional (3D) models [14]. The most common method for practicing CI surgery involves the use of a cadaver/temporal bone. Though this method of training is time-honored, it has many limitations. Cadaveric temporal bones have become highly regulated, more expensive, and less accessible to trainees. Also, preservation chemicals lead to special equipment that is needed to protect the surgeons from the risk of chemical exposure. These limitations make it difficult for surgeons to gain extensive experience in CI surgery. VR is another method that allows any surgical task to be replicated in order to facilitate education and assessment, giving the learner the opportunity to repeatedly try new skills, make mistakes, and allow the surgeon competency to evolve with more training. Studies have shown that VR-trained surgeons complete procedures 20% faster than traditionally trained surgeons and complete 38% more of the procedure steps correctly [14,15]. However, the problems associated with virtual reality include the high cost and amount of resources needed to create a high-quality VR system, a learning curve to overcome using the VR equipment, and sometimes nausea and motion sickness. Current VR systems lack controllers that enable fine motor skills and accurate tactile feedback during simulation that accurately mimics hand motions, and VR training specifically for CI surgery is not widely available.

Three-dimensional models can be used to practice CI surgery. There are many different types of 3D models that can be used in surgical training such as anatomical models (i.e., TruCorp, Lurgan, Craigavon, UK), plastinated specimens, and 3D-printed models. Of these types, the 3D-printed models provide students with the most realistic tactile learning experience [16]. In general, 3D-printed models involve reconstruction and modeling techniques where a digital (cad, ct reconstruction, etc.) file is uploaded to a 3D printer, which then prints an anatomically precise 3D object. The technology is also known as “additive manufacturing” because it adds layer upon layer of material (thermoplastics, photopolymers, human cells and gelatin, epoxy resins, metals, etc.) to build an object. The primary shortfall of current 3D-printed models relative to CI surgery training is that they lack detailed real-time feedback.

To fully understand the feedback requirements of the 3D model, a detailed understanding of the surgical steps involved is necessary. As was mentioned previously, the surgeon completes a mastoidectomy by removing the hollow bone behind the ear [17]. The mastoid bone surface is exposed, and the honeycomb partitions of the mastoid bone are drilled down to where they connect to the middle ear. The surgeon's goal is to locate the round window of the cochlea. Having exposed the round window, the surgeon then drills a well and channels into the skull to seat the implant. The round window is carefully opened, and the implanted electrode is then inserted through the round window opening. The electrode is then threaded into the cochlea as far as it will go. During electrode threading, the surgeon is unable to see past the round window. The surgeon must learn to blindly use the walls of the cochlea to guide the thin electrode into position and avoid several issues that can occur. If too much pressure is used due to the electrode being inserted too fast, intracochlear damage can occur or the electrode can be damaged. Electrode kinking can occur if the electrode is not traveling along the wall correctly. Electrode tip fold-over can occur if the electrode snags and continues to be fed into the cochlea. Electrode backout can occur if the electrode insertion depth is insufficient for the cochlea geometry to hold it in place. In addition, some of the electrode tips have a hook shape that has an ideal orientation relative to the auditory nerve, however, the surgeon is currently unable to assess final electrode tip orientation due to the lack of visual feedback. Spinal fluid may leak through the round window opening and needs to be sealed at the conclusion of the electrode insertion, the round window opening is sealed with tissue from the patient.

With this knowledge, a smart, realistic, 3D-printed model of the cochlea was developed in this project that provides visual feedback using a video camera and a customized graphical user interface (GUI). The model was designed to:

  1. (1)

    Provide real-time feedback for insertion technique improvement during practice

    • Tactile feedback that allows the trainee to realistically feel an electrode inserted into an actual-sized cochlea replica, filled with fluid, via a round window membrane.

    • Visual feedback that allows the trainee to observe real-time conditions such as electrode kinking and back-out, tip fold-over and orientation, and overall insertion speed and depth.

  2. (2)

    Provide tracking feedback for surgical method technique improvement over time

    • Tracking feedback that records for each trainee session the insertion speed and depth, plus the number of electrode kinks, fold-overs, and back-outs, indicating if acceptable results were obtained.

  3. (3)

    Be portable, reusable, and economical.

Model Design

Hardware.

To 3D-print a realistic cochlea, a relevant digital file was needed to upload to a 3D printer. A realistic cochlea is hollow, filled with spinal fluid, and has a round window membrane covering its entrance. Digital files with this level of detail were available via reconstruction images established from scans performed on actual patients, such as from a magnetic resonance image (MRI). Figure 1 displays the SolidWorks drawing developed for 3D printing of the cochlea. To print the 3D cochlea, the print material needed to be compatible with water, as cerebrospinal fluid typically fills the cochlea canal. Cerebrospinal fluid is 99% water, and from a tactile point of view feels like water, hence it was used as an appropriate replica [18]. The material also needed to be transparent to allow for the visualization inside the cochlea canal. Finally, the material needed to be sturdy to allow for repetitive use as a training tool. The material selected was Vero Clear and was 3D printed on a Stratasys Object machine. Figure 2 shows an image of the 3D-printed cochlea. When printed in a horizontal position, two artifacts were present that were later found to interfere with electrode detection. When printed in the vertical position only one artifact was present, and it was largely out of the electrode viewing area, so this was the method of choice.

Fig. 1
Customized SolidWorks drawing of the cochlea
Fig. 1
Customized SolidWorks drawing of the cochlea
Close modal
Fig. 2
3D-printed cochlea
Fig. 2
3D-printed cochlea
Close modal

A reusable electrode, the Nucleus 24 Contour Advance Practice Electrode (see Fig. 3) was selected for the training purposes. To enhance the contrast needed for real-time visual feedback, the electrode was colored black (see Fig. 4) so it could be easily seen inside the 3D-printed cochlea by a camera. Many sensing techniques were initially explored for this project to determine which would provide the best real-time feedback during electrode insertion. Capacitance changes in the electrode were explored for detecting electrode location. Resistance and impedance changes in the electrode were explored for detecting kinks and fold-overs by adapting an active insertion monitoring (AIM) system that is sometimes employed on the electrode during CI surgery [19]. All of these techniques were abandoned in favor of video/ray analysis which allowed the detection of all the electrode parameters at the same time. Video/ray analysis involves using a video camera to view and record the electrode insertion while it occurs inside the transparent 3D-printed cochlea and analyzing that live video stream with ray analysis software that computes real-time electrode depth, speed, orientation, kinks, fold-overs, and back-outs. Using a graphical user interface (GUI), real-time information could be displayed to the trainee, and training data could be recorded and tracked over time. For both visual feedback and flexibility, it was desired that the video camera would produce a high-resolution, live feed to a computer via a USB port, to allow for a computer monitor display and software manipulation. The size of the camera needed to be comparable to the size of the 3D-printed cochlea. The camera ultimately meeting these specifications was an Omnivision wide angle ¼” mini video web camera shown in Fig. 5.

Fig. 3
Nucleus 24 contour advance practice electrode
Fig. 3
Nucleus 24 contour advance practice electrode
Close modal
Fig. 4
Colored electrode inserted in the 3D-printed cochlea
Fig. 4
Colored electrode inserted in the 3D-printed cochlea
Close modal
Fig. 5
Omnivision wide angle ¼” mini video web camera
Fig. 5
Omnivision wide angle ¼” mini video web camera
Close modal

Additional mounting pieces were designed and printed to complete the 3D model: a funnel, a removable funnel insert, a funnel pin, and a camera/lighting mount. An access funnel was used to mimic the distance from the outside of the temporal bone to the round window inside the skull (30 mm as measured in a cadaver). This funnel was also designed to hold a membrane in place to mimic the round window at the entrance to the cochlea. Replaceable plastic wrap was used for this membrane in the final version of the setup that was held in place by a removable funnel insert. CI surgeons and trainees found this representation acceptable in the validation trials.

Figure 6 shows the final prototype setup with the funnel attached to the 3D-printed cochlea model, the camera on inside, the light on the other side, the customized mount that holds all the components, and customized GUI for training the surgeons in CI electrode insertion. A Davis and Sanford Vista Explorer 60” tripod was selected to allow for the vertical and horizontal flexibility to mimic optimal operating table location for each surgeon performing surgery on either ear.

Fig. 6

Figure 7 illustrates the details for the customized 3D-printed mount. A camera/lighting mount was designed to place the cochlea model at a surgical table level and to hold the web camera. A camera holder and camera hole were designed to snugly hold the web camera in position on either side of the cochlea and to allow the camera lens to fit through the hole at the focus depth specified to achieve clear images of the electrode inside the cochlea. A 3-watt light bulb was positioned at the hole opposite the camera and was powered by an AA battery.

Fig. 7
3D-printed components of complete design
Fig. 7
3D-printed components of complete design
Close modal

Software/Graphical User Interface.

A software platform was created to provide visual and tracking feedback via a GUI on a computer monitor. Real-time visual feedback was desired that allowed the trainee to observe conditions such as electrode kinking and back-out, tip fold-over and orientation, and electrode insertion speed and depth while inserting the electrode into the 3D-printed cochlea. Tracking feedback was desired to create records over time so that each trainee could demonstrate improvement in the electrode insertion technique with and without visual feedback. The tracking feedback desired for each trainee was insertion speed and depth, along with the number of electrode kinks, fold-overs, and back-outs. The code was written to interact with the live-stream video camera in two ways. First, the code was developed to display the live video stream of the 3D cochlea in the GUI. This was done so that the trainee could observe the electrode tip conditions in real time, such as tip orientation. Second, additional code was added to take a series of images from the live stream video so that real-time analysis could be done relative to the electrode insertion technique. The image analysis indicates real-time insertion depth, speed, kinks, fold-overs, or back-outs in real-time. Figure 8 shows the full GUI the user sees when using the training setup. The upper left image shows the real-time image stream of the actual electrode insertion. The image next to it displays the live tracking of the electrode tip for further analysis.

Fig. 8
Full GUI display

Figure 9 shows the cochlea as seen by the camera, and Fig. 10 shows the initial setup for the GUI just before a training session. The program has the user select the center point of the cochlea, the number of rays desired, and the start/stop angles for the rays. Typically, one starts with a low number of rays (say 15), and then increases them to get more resolution. A higher number of rays causes the update rate of the video to slow down, so one can find the right balance between resolution and update speed on the live feedback.

Fig. 9
Typical initial set-up file image
Fig. 9
Typical initial set-up file image
Close modal
Fig. 10
To create set-up file
Fig. 10
To create set-up file
Close modal

Once the “Start” button is pushed in the GUI, numerous customized functions are used in the initialization to develop a look-up table (LUT) and read the ray segment information. First, it calculates a Gaussian distribution and stores it in the LUT for use in enhancing the white pixel values and minimizing the black pixel values.

The program was designed to process images during electrode insertion. The program displays the current image, finds the electrode tip position, and stores this data. This is the image displayed in the upper left of the GUI. The total electrode thickness value is stored and is later compared to the set thickness of a normal electrode. If it is greater than the set thickness, then a kink or fold-over has occurred. Figure 11 reveals a sample tip fold-over during electrode insertion in the cochlea.

Fig. 11
A tip fold-over during electrode insertion
Fig. 11
A tip fold-over during electrode insertion
Close modal

The change in distance divided by the change in time is used to calculate the instantaneous speed (in pixels/sec) of electrode insertion. No digital filters were required here as the noise level was minimum. The calculated speed value can be positive or negative; if the electrode is inserted further into the cochlea, it is positive, if the electrode backs out, it is negative. Next, the velocity angle is calculated and used in displaying the red ray where the electrode tip should be at, at that time. Kinks and fold-overs are determined by comparing the values for the electrode thickness with the known electrode thickness. If the thickness is greater, the kink-count variable is incremented by 1, indicating a kink or fold-over has occurred. Back-outs are detected when the instantaneous speed is found to be negative and greater than 2 pixels/sec in magnitude. Smaller magnitudes were determined to be noise. When a back-out is detected, the number of back-outs is increased by 1. The program is designed to plot the real-time depth of insertion (pixels) and speed of insertion (pixels/sec) of the electrode as it loops through each image throughout the insertion process. The depth and speed can be converted to mm and mm/s by using the conversion of 25 mm = 380 pixels for all work done in this paper.

Acceptable values for insertion depth and speed are indicated with solid-colored lines and can be changed by anyone overseeing the training. The physical electrode has a blue line on it that indicates the appropriate insertion depth in the cochlea. The acceptable instantaneous insertion speed range was established by recording and analyzing practicing surgeons' insertion technique repeatedly when no kinks, fold-overs, or back-outs occurred. This acceptable insertion speed is displayed throughout the insertion process to the trainee as solid green and red lines in the lower graph. The trainee's insertion speed is tracked in light blue relative to the green and red lines. Note that both positive and negative values are graphed. By observing the two graphs shown on the left side of the GUI during insertion, the trainee can learn to insert the electrode at a continuously acceptable speed and to an acceptable depth. Note that the real-time occurrence of the number of kinks and back-outs is tracked as well in the right portion of the GUI. It is undesirable to have any of these occur during electrode insertion. By observing when these occur, the trainee learns what insertion techniques to avoid. The total number of back-outs and kinks (including fold-overs) are recorded for tracking purposes as well as to evaluate electrode insertion technique improvement.

Test Trials and Results

Four sets of test trials were conducted at Mercy Health Hospital in Youngstown, Ohio, and at the University of Akron. The first three trials lead to significant improvements in the design, and the final and fourth trial demonstrated that the device helped significantly in the training leading to improved outcomes. The three trials at the hospital were part of an ongoing anatomy lab for medical residents, fellows, and early-career surgeons. There were several stations for trainees to rotate through during these labs. For the three trials in the hospital one station was devoted to practicing CI electrode insertion on the developed 3D-printed cochlear model presented in this paper (see Fig. 12).

Fig. 12
Trainee using the prototype at clinical trials
Fig. 12
Trainee using the prototype at clinical trials
Close modal

A practicing clinician with over twenty years of experience instructed the trainees on the optimal electrode insertion process. As mentioned before, four metrics were selected in the training process and all trainees were instructed on the importance of each:

  1. Depth (D) → ideal depth was 25 mm (indicated by the blue line on the electrode)

  2. Insertion speed (IS) → lower the better

  3. Number of back-outs (BO) → lower the better

  4. Number of kinks, fold overs (K/FO) → lower the better

Trial 1.

During the first trial, the training sessions were filmed with the video camera so that the software and GUI could be developed. After the trial, the training session data was analyzed to establish benchmarks for acceptable electrode insertion and verify that there was a clear difference in technique between the trainees and the experienced surgeon. In general, it was found that the trainees' insertion speeds (average and maximum) were greater than the CI surgeon's insertion speeds. Also, 75% of the trainees averaged higher electrode back-outs than the CI surgeon. These results illustrated the variation in trainee insertion techniques that would benefit from repetitive, real-time, tactile, and GUI feedback that indicated acceptable levels for each parameter measured while using the 3D-printed cochlear model.

Several adjustments were made to the 3D-printed model after this trial. For instance, the cochlea originally was separated from the funnel, but since it tended to move during use and a lip was felt at the funnel/cochlea transition, it was later printed as one continuous piece. Also, the original prototype did not include a funnel insert and membrane at the cochlear entrance. This was later added for better tactile feedback. Relative to the camera, it was found that both the focal length and lighting of the electrode needed optimization from its original settings. These were adjusted by changing the camera mount location and by adding lighting to the setup. The mount tended to move and was unsymmetrical, leading to setup issues with reuse. Notches and pins were designed to assure ease of setup and sturdiness with each training session. Finally, it was found that a vertical print orientation for the cochlea resulted in less disruptive exterior artifact material than horizontal print orientation.

Trial 2.

The primary focus of this trial was to assess the practicality of the GUI to provide useful feedback to the trainee during electrode insertion. In general, the live feed and real-time parameter feedback were found to be useful. It was determined that the speed of electrode insertion was a key metric in technique improvement for the trainees. Several improvements for the GUI were realized at this second clinical trial. As expected, it was found that the calibration of the depth and speed indications were needed, and acceptable ranges needed to be clearly delineated for real-time benefit. It was during this session that a “stop” button was found to be useful when a test was not progressing as desired. Also, image processing was slower than desired. This led to a focused effort to streamline the code for more efficient image processing.

Trial 3.

After the modifications were made to the GUI, an interim trial was conducted with six engineering students at The University of Akron. The primary focus of this trial was to demonstrate that the real-time GUI feedback was sufficient to cause even a novice to improve electrode insertion technique with repetitive use. Each student initially inserted the electrode to the blue line without the use of the GUI several times. Then each student repetitively practiced inserting the electrode while using the GUI for real-time feedback. Then the students tried inserting the electrode without the visual feedback. The results of this interim trial are shown below in Table 1. The Before column is the average of results without visual feedback, and the After column is the average of the results after training, but without visual feedback while inserting the electrodes.

Table 1

Trial 3 results

BeforeAfter% change
Depth22.5224.18mm7.35%
Speed4.313.05mm/s−29.09%
No of BO3.571.79−49.81%
No of K/FO2.171.40−35.30%
BeforeAfter% change
Depth22.5224.18mm7.35%
Speed4.313.05mm/s−29.09%
No of BO3.571.79−49.81%
No of K/FO2.171.40−35.30%

Trial 3 was the first time that consistent data was generated in all four metrics and was analyzed. Figure 13 shows the results. The novice engineering students showed an overall improvement in all four metrics, once they were trained on the hardware/software.

Fig. 13

D: 7.35% improvement in depth

IS: 29.09% reduction in insertion speed

BO: 49.81% reduction in back-outs

K/FO: 35.30% reduction in kinks/fold-overs

Trial 4.

The final clinical trial was held on 4/1/2022 at the hospital using the complete prototype. The primary focus of this trial was to document improvement in trainee electrode insertion techniques after being trained by the prototype. Each of the seven trainees (medical students/residents) inserted the electrode at least five times without the use of the GUI (before) and then at least five times after they had practiced with the setup and GUI (after). The results of this trial showed an improvement in electrode insertion technique. The results of this final trial are shown below.

Trial 4 results were consistent with the trial 3 results, but the trainees were all medical students/residents with minimal cochlear insertion training, and the results are tabulated in Table 2. Figure 14 shows the results. As in trial 3, the medical trainees showed an overall improvement in all four metrics, once they were trained on the hardware/software.

Fig. 14
Table 2

Trial 4 results

BeforeAfter% change
Depth21.5224.58mm14.20%
Speed5.423.02mm/s−44.24%
No of BO0.910.43−52.90%
No of K/FO0.460.16−64.89%
BeforeAfter% change
Depth21.5224.58mm14.20%
Speed5.423.02mm/s−44.24%
No of BO0.910.43−52.90%
No of K/FO0.460.16−64.89%

D: 14.20% improvement in depth

IS: 44.24% reduction is insertion speed

BO: 52.90% reduction in back-outs

K/FO: 64.89% reduction in kinks/fold-overs

Clinical Implications and Limitations

This work was inspired by the lack of consistent training models for novice medical trainees. Virtual reality models do not provide comprehensive training experience and cadaver model training is limited in number. The proposed platform provides a clinically relevant experience but is no substitute for actual experience. Self-learning through practice provides confidence so that the medical trainees can move forward with more skill sets. The proposed platform does not have the temporal bone drilling experience, as that remains as future work. Another limitation is the programing which is currently done using commercial a software package (Matlab by Mathworks) and a compiled version is provided to the user. Future work entails converting the software to an open-source software using Python.

Conclusions

A realistic 3D-printed model of the cochlea was developed that provides useful feedback using a mini-web video camera and a GUI that employs custom software with ray analysis as its backbone. The model provides real-time feedback for CI electrode insertion technique improvement during practice by providing both tactile feedback and real-time GUI visual feedback that shows electrode depth, speed of insertion, back outs, and kinks/fold-overs. The model was tested with multiple trials. In the final trial conducted for medical students/residents, there was an overall improvement in all four metrics after they were trained on the hardware/software. There was a 14.20% improvement in insertion depth, 44.24% reduction in insertion speed, 52.90% reduction in back-outs, and a 64.89% reduction in kinks/fold-overs. The advantage of this model is that surgeons can use it as many times as they like, as the whole setup is easy, economical, and reusable. The test trials have demonstrated that repeated use of the 3D-printed model and the GUI help improve overall CI procedure outcomes for ENT surgeon trainees. Advantages of this model include that it may be used as many times as desired and the whole setup is easy, portable, reusable, and economical.

References

1.
Fitzpatrick
,
E. M.
,
Olds
,
J.
,
Gaboury
,
I.
,
McCrae
,
R.
,
Schramm
,
D.
, and
Durieux-Smith
,
A.
,
2012
, “
Comparison of Outcomes in Children With Hearing Aids and Cochlear Implants
,”
Cochlear Implants Int.
,
13
(
1
), pp.
5
15
.10.1179/146701011X12950038111611
2.
Dettman
,
S. J.
,
D'Costa
,
W. A.
,
Dowell
,
R. C.
,
Winton
,
S. J.
,
Hill
,
K. L.
, and
Williams
,
S. S.
,
2004
, “
Cochlear Implants for Children With Significant Residual Hearing
,”
Arch. Otolaryngology – Head Neck Surg.
,
130
(
5
), pp.
612
618
.10.1001/archotol.130.5.612
3.
Types, Causes and Treatment
,
2022
, “Hearing Loss Association of America (HLAA), Conductive Hearing Loss,” HLAA, Rockville, MD, accessed Dec. 10, 2021, https://www.hearingloss.org/hearing-help/hearing-loss-basics/types-causes-and-treatment/
4.
Loizou
,
P. C.
,
1999
, “
Introduction to Cochlear Implants
,”
IEEE Eng. Med. Biol. Mag.
,
18
(
1
), pp.
32
42
.10.1109/51.740962
5.
Mayo Clinic ENT Department
,
2022
,
Cochlear Implants
,
Mayo Clinic
, Scottsdale, AZ, accessed Dec. 20, 2022, https://www.mayoclinic.org/tests-procedures/cochlear-implants/about/pac-20385021
6.
Farinetti
,
A.
,
Mancini
,
J.
,
Ben Gharbia
,
D.
,
Roman
,
R.
,
Nicollas
,
S.
, and
Triglia
,
J. M.
,
2014
, “
Cochlear Implant Complications in 403 Patients: Comparative Study of Adults and Children and Review of the Literature
,”
Eur. Ann. Otorhinolaryngol. Head Neck Dis.
,
131
(
3
), pp.
177
182
.10.1016/j.anorl.2013.05.005
7.
Cohen
,
N. L.
, and
Hoffman
,
R. A.
,
1991
, “
Complications of Cochlear Implant Surgery in Adults and Children
,”
Ann. Ontol. Rhinol. Laryngol.
,
100
(
9
), pp.
708
711
.10.1177/000348949110000903
8.
Theunisse
,
H. J.
,
Pennings
,
R. J. E.
,
Kunst
,
H. P. M.
,
Mulder
,
J. J.
, and
Mylanus
,
E. A. M.
,
2018
, “
Risk Factors for Complications in Cochlear Implant Surgery
,”
Eur. Arch. Otorhinolaryngol.
,
275
(
4
), pp.
895
903
.10.1007/s00405-018-4901-z
9.
Gheorghe
,
D.
,
2015
, “
Complications in Cochlear Implant Surgery
,”
J. Med. Life
, 8(3), pp.
329
332
.https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4556914/
10.
Ishiyama
,
A.
,
Risi
,
F.
, and
Boyd
,
P.
,
2020
, “
Potential Insertion Complications With Cochlear Implant Electrodes
,”
Cochlear Implants Int.
,
21
(
4
), pp.
206
219
.10.1080/14670100.2020.1730066
11.
Raghunandhan
,
S.
,
Kameswaran
,
M.
,
Anand Kumar
,
R. S.
,
Agarwal
,
A. K.
, and
Hossain
,
M. D.
,
2014
, “
A Study of Complications and Morbidity Profile in Cochlear Implantation: The MERF Experience
,”
Indian J. Otolaryngol. Head Neck Surg.
,
66
(
S1
), pp.
161
168
.10.1007/s12070-011-0387-3
12.
Mertens
,
G.
,
Van Rompaey
,
V.
,
Van de Heyning
,
P.
,
Gorris
,
E.
, and
Topsakal
,
V.
,
2020
, “
Prediction of the Cochlear Implant Electrode Insertion Depth: Clinical Applicability of Two Analytical Cochlear Models
,”
Sci. Rep.
,
10
(
1
), p.
3340
.10.1038/s41598-020-58648-6
13.
Torres
,
R.
,
2016
, “
Variability of the Mental Representation of the Cochlear Anatomy During Cochlear Implantation
,”
Eur. Arch. Otorhinolaryngol.
,
273
(
8
), pp.
2009
2018
14.
Blumstein
,
G.
,
2019
, “
Research: How Virtual Reality Can Help Train Surgeons
,
Harvard Business Review
, accessed Dec. 1, 2021, https://hbr.org/2019/10/research-how-virtual-reality-can-help-train-surgeons
15.
Surgical Science
,
2023
, “
Training That Makes a Difference, Surgical Science
,” Göteborg, Sweden, accessed Nov. 20, 2023, https://surgicalscience.com/simulators/?gclid=Cj0KCQiApOyqBhDlARIsAGfnyMpmRN4yDDtgU8rsv54I2pKPApxnNDgwb-Yvm6P_emQpAgLz0QdLefQaAoTeEALw_wcB
16.
Bati
,
A.
,
2020
, “
3D Modelling for Realistic Training and Learning
,”
Turk. J. Biochem.
, 47(2), pp.
179
181
.10.1515/tjb-2019-0182
17.
Roland
,
J.
,
2005
,
Cochlear Implant Electrode Insertion
,
Science Direct
, Elsevier, Amsterdam, The Netherlands, accessed Oct. 20, 2019, https://www.sciencedirect.com/science/article/abs/pii/S1043181005000138
18.
Bateman
,
J.
,
2014
,
Manufacture of a Prototype for Functionality Studies of Cerebrospinal Fluid
,
National Science Foundation
, Alexandria, VA, accessed, Nov. 14, 2021, http://msoe.s3.amazonaws.com/files/resources/jen-bateman-paper.pdf
19.
Duarte
,
O.
, Sonova,
2019
,
Advanced Bionics Launches Active Insertion Monitoring (AIM) System: An innovative Monitoring Solution for Implant Surgery and Post-op
,
Sonova
, Stäfa, Switzerland, accessed Dec. 12, 2021, https://www.sonova.com/en/media/advanced-bionics-launches-active-insertion-monitoring-aim-system-innovative-monitoring