Neural Correlates of Loudness Coding in Two Types of Cochlear Implants—A Model Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript "Neural Correlates of Loudness Coding in Two Types of Cochlear Implants – A Model Study" presents an innovative computational description of two cochlear implant speech coding strategies, ACE (Cochlear) and F120 (Advanced Bionics). By simulating both systems in a "virtual cochlea and nerve" model, the study effectively evaluates the effects of the two strategies on virtual loudness encoding, offering insights that are difficult to obtain from patient-based comparisons.

The methodology is innovative, the clinical relevance is evident, and the results presentation is clear and concise. 

 

However, I would like to highlight three aspects of the study design that require revision or clarification:

1. It is unclear why the authors chose a lateral wall electrode for Cochlear and a pre-curved electrode for Advanced Bionics, given that their portfolio includes two lateral wall devices. 

2. Description of Acoustic Processing for ACE

The acoustic pre-processing pipeline for F120 is documented: the authors used the Advanced Bionics Generic-Python-Toolbox, with scaling (audiomixer full scale of 109.6 dB), a defined dynamic range (40 dB), AGC and pre-emphasis enabled, and noise reduction disabled.

In contrast, for ACE, the description is limited to stating that the Nucleus MATLAB Toolbox was used. The authors then refer to default values for T-level (33.86 dB) and C-level (65.35 dB), without mentioning the front-end signal conditioning or whether ADRO and ASC were activated in the back-end. This lack of detail makes it difficult to assess whether the two strategies were modeled under acoustically equivalent conditions, and it raises questions about the source of observed differences in neural activation. A complete and symmetrical description of the acoustic processing pipelines for both strategies is recommended.

3. Stimulation Rate Choice for ACE

The paper compares ACE at 900 pps/channel with F120 at 1850 pps/channel, but does not justify why higher ACE rates, readily available in clinical fitting (e.g., 1800 pps), were not considered.

Two points should be noted:

The reference to Balkany et al. is used to justify the 900 pps rate. Still, their findings highlight variability in patient preference and performance, rather than supporting 900 pps as an optimal or exclusive choice, even if the majority prefers it.

The ACE strategy is technically capable of operating at 1800 pps, which would provide a more equitable comparison with F120. Failing to match or vary the pulse rates limits the validity of the conclusions about loudness encoding differences.

 

Recommendations

The following recommendations are suggested.

Provide a complete description of the ACE acoustic preprocessing pipeline, including all relevant settings.

Include a comparative discussion of the acoustic preprocessing steps for both strategies, focusing on their potential impact on loudness encoding.

Either conduct simulations at matched stimulation rates (e.g., 1800 pps for ACE) or, at minimum, discuss how differences in rate affect neural activation and the interpretation of results.

Finally, explain why the authors compared two electrodes with different intra-cochlear positioning.

These additions would substantially strengthen the paper's validity and applicability.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting approach to use an established cochlear nerve model to compare different cochlear implants applying different coding strategies. Here the parameter loudness is of main interest.

35: In typical hearing >> please specify clearer

45: changing the pulse shape >> specify in more detail differences between pulsatile and analogue stimulation

Methods: why you did compare a lateral wall electrode against mid-scala electrode. A modiolus hugging electrode for ACE would be more appropriate.

In Fig 4 the total spike rate is described. The temporal fine structure in Fig 3 is hard to catch. Are there differences in rate depending on distance to the electrode and how is that represented in your model?

325: . N-of-M strategies stimulate the (in this case) eight most dominant analysis bands >> are these ever 8 bands?

331: ACE starts activating spikes at 25 dB. This may be an effect of the SCS as you describe. But how about neural effects? How about facilitation – is this incorporated in your model?
Here a single electrode was stimulated on. In normal use the overall stimulation rate is much higher and would code sound due to facilitation and refractoriness.
>> describe whether such characteristics are part of your model.

Please give an out look for more complex sound – like temporal coding – using your model.

Author Response

Please see the attachment. 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors revised the paper correctly following my requests 

Reviewer 2 Report

Comments and Suggestions for Authors

The changes support your paper to convey a clearer message.