Does the Remote Microphone Still Outperform the Pre-Processing Algorithms? A Group Study in Adult Nucleus Recipients

Round 1

Reviewer 1 Report

The study has some shortcomings as the authors acknowledge. Mainly, the use of the S0N0 condition prevents the possibly advantageous activation of directional microphone processing but also diminishes possible advantages of the remote mic.

The reasons why the remote mic was superior to the SCAN processing remain obscure. One possibility is the condition of the test room. This should be mentioned in the methods section (reverberation time, size of the room, additional noise sources). Another possibility for the difference between remote mic and processor signal pickup might be the reduction of higher frequencies in the spectral distribution with increasing distance from source to receiver. Reduced gain for higher frequencies might affect consonant recognition and therefore the remote mic condition might be favorable. Another possibility might be differences in gain, i.e. different loudness mapping for the two conditions. One way of evaluating these acoustic parameters would be the analysis of electrodograms for the two conditions (e.g. noise only presentation). Some of these possible factors should be discussed. A more detailed description of the signal processing with SCAN for the S0N0 condition might also be helpful. There is no beamforming activated presumably and the environment remains unchanged. Thus, SCAN will not change the processing mode. However, there might be different AGC or ASC settings compared with the remote mic condition.

Statistical analysis: In lines 138 to 143 the SRT results are presented. Here, an ANOVA might be appropriate to verify that the processor type did not play a role.

Matrix test: L76: to avoid training effect -> this sentence seems misleading. The matrix test requires training lists since results usually stabilize only after the first two or three lists. Thereafter, there is no or only a minimal training effect. However, when too many conditions are tested in sequence the subjects might get tired and results become unreliable again. Thus, this might be described as a fatigue effect which does not seem to be relevant in this investigation.

Title: Remote Microphone (singular), study in adult Nucleus recipients (not: adults)

L17: more effective. (omit: in overcome it)

L18: recipients

L42: speaking to her (refers to the person)

L48: processors, manufacturers

L125: subjects

L243: Finally, while reverberation...

L251: in contrast

L260: Wolfe et al.'s findings

L265: what was assumed by Wolfe et al.

Author Response

We want to thank the reviewer for the kind work.

Its suggestions will make our paper stronger. 

Point-by-point responses will follow.

The study has some shortcomings as the authors acknowledge. Mainly, the use of the S0N0 condition prevents the possibly advantageous activation of directional microphone processing but also diminishes possible advantages of the remote mic.

We agree with the reviewer that the S0N0 condition partially conditioned the study. We discussed that in the "Limit of the study" paragraph.

The reasons why the remote mic was superior to the SCAN processing remain obscure. One possibility is the condition of the test room. This should be mentioned in the methods section (reverberation time, size of the room, additional noise sources).

In the methods section we added a paragraph regarding the testing environment. All the subjects were tested in an acoustically isolated room, with sound-absorbing panel on the walls and on the roof (see lines 92-94).   

Another possibility for the difference between remote mic and processor signal pickup might be the reduction of higher frequencies in the spectral distribution with increasing distance from source to receiver. Reduced gain for higher frequencies might affect consonant recognition and therefore the remote mic condition might be favorable. Another possibility might be differences in gain, i.e. different loudness mapping for the two conditions. One way of evaluating these acoustic parameters would be the analysis of electrodograms for the two conditions (e.g. noise only presentation). Some of these possible factors should be discussed.

We thank the reviewer for the kind explanations. Those are very reasonable explanation and we provided a discussion on this (see lines 285-300).

Unfortunately, we are not able to produce the electrodograms for this study, but the reviewer idea can provide a hint for future studies. 

A more detailed description of the signal processing with SCAN for the S0N0 condition might also be helpful. There is no beamforming activated presumably and the environment remains unchanged. Thus, SCAN will not change the processing mode. However, there might be different AGC or ASC settings compared with the remote mic condition.

A whole paragraph regarding SCAN functioning, also in S0N0 configuration, had been added (see lines 215 - 226).

Statistical analysis: In lines 138 to 143 the SRT results are presented. Here, an ANOVA might be appropriate to verify that the processor type did not play a role.

We double checked the statistical results and we run the ANOVA for the comparison of SRT in the different processor groups. We can confirm that there are not statistically significant differences.   

Comments on the Quality of English Language

An extensive English language editing has been provided, also according to reviewer's kind suggestions. 

Reviewer 2 Report

Dear Editor

I have read the Manuscript entitled: "Does the Remote Microphones still outperform the Preprocessing Alogarithms? A group study in adults Nucleus recipients" submitted to the Otorhinolaryngology, Hearing and Balance Medicine. The topic is definitely relevant. The findings of this study can be of interest to the readers and may have implications for professionals in the clinic.

However, the manuscript needs some revision before it is ready for publication.

The English needs some corrections: for example, line 17: in overcome it, or line 86: for more than 20 different languages etc.

In the Introduction there is need to explain the difference between the preprocessing SCAN and the wireless MiniMic. This is elaborated in the discussion, but should appear already in the introduction.

In the Method section you should mention the ethics committee approval.

Also, was the SRT measured always first with the SCAN and later with the MiniMic? Maybe there was an effect of training? It is unclear when you compare lines 75-76 to lines 113-119

Line 65: what kind of noise was used? Babble? What dB was used? HL or SPL?

Line 120: if I understand correct this is a dynamic test that adjusts the SNR according to the previous sentence response. How is it that all subjects repeated 20 sentences?

Throughout the Results section the terms SRT and SNR are used, there is a need to differentiate them in a clear manner.

Line 135: the use of Delta is described, but not clear for what purpose until line 145. Please change order

Discussion line 222: please mention references.

On page 7 you mention the comparison between adaptive remote microphones such as Roger and the MiniMic. Please elaborate on the advantages of each (any reasons why in Razza's study better performance with fixed MiniMic compared to Roger?)

A very similar study to yours was published by De Ceulaer in 2017.  It should be mentioned, and need to further clarify your study's addition

The use of cochlear's SCAN and wireless microphones to improve speech understanding in noise with the Nucleus6® CP900 processor.

De Ceulaer G, Pascoal D, Vanpoucke F, Govaerts PJ.Int J Audiol. 2017 Nov;56(11):837-843.

need an expert to review the English

Author Response

I have read the Manuscript entitled: "Does the Remote Microphones still outperform the Preprocessing Alogarithms? A group study in adults Nucleus recipients" submitted to the Otorhinolaryngology, Hearing and Balance Medicine. The topic is definitely relevant. The findings of this study can be of interest to the readers and may have implications for professionals in the clinic.

However, the manuscript needs some revision before it is ready for publication.

We would like to thank the reviewer for its kind appraisal and its key suggestions.

Point-by-point responses will follow. 

The English needs some corrections: for example, line 17: in overcome it, or line 86: for more than 20 different languages etc.

The document underwent a deep English language revision, also according to your comments.  

In the Introduction there is need to explain the difference between the preprocessing SCAN and the wireless MiniMic. This is elaborated in the discussion, but should appear already in the introduction.

We provided a brief explanation of the difference in the introduction, according to the reviewer suggestions (see lines 63-68).

In the Method section you should mention the ethics committee approval.

Done (see lines 141-142).

Also, was the SRT measured always first with the SCAN and later with the MiniMic? Maybe there was an effect of training? It is unclear when you compare lines 75-76 to lines 113-119

After two or three lists for practicing, some of the patients did the test firstly with SCAN and others with the MiniMic, randomly. That has been done to avoid bias in results evaluation (e.g. if all the subjects would have done the test always with SCAN or MiniMic first, we could have speculated that the results was due to a training effect or, viceversa, due to tiredness effect).

We provided an editing on lines 84-85 and 122-125 for a better explanation of the concept.   

Line 65: what kind of noise was used? Babble? What dB was used? HL or SPL?

In OLSA Matrix test, the standard noise is a typical babble noise and it is constantly presented at 65dB HL.

Line 120: if I understand correct this is a dynamic test that adjusts the SNR according to the previous sentence response. How is it that all subjects repeated 20 sentences?

That is absolutely right. OLSA Matrix test is an adaptive test that adapt its difficulty (in term of SRT) according to subject responses. It is normal in this test to understand sentences partially. We provided a better explanation of "all the subject repeated all the 20 sentences" writing that the 20 sentences were administered to the patients. 

Throughout the Results section the terms SRT and SNR are used, there is a need to differentiate them in a clear manner.

We explain that at the beginning of the discussion section.

Line 135: the use of Delta is described, but not clear for what purpose until line 145. Please change order

Done (see lines 133-135).

On page 7 you mention the comparison between adaptive remote microphones such as Roger and the MiniMic. Please elaborate on the advantages of each (any reasons why in Razza's study better performance with fixed MiniMic compared to Roger?)

Done (see lines 320-328).

A very similar study to yours was published by De Ceulaer in 2017.  It should be mentioned, and need to further clarify your study's addition

We thank the reviewer for the report. We managed to integrate the article in our manuscript (see mostly lines 331-341).

Comments on the Quality of English Language   As we said before, a broad English Language editing has been done.   

Round 2

Reviewer 1 Report

The revision of the manuscript has taken into account all my previous comments. I appreciate the additional information about the test setup and environment, the additional ANOVA statistical analyses and the updated discussion section. I have no further comments.

Author Response

We thank the reviewer for his fine revision.

Your suggestions definitively made our paper better.

Reviewer 2 Report

although the authors have implemented important changes, there are still missing details in the methods section. They describe a relative test that finds the SNR where 50% is understood. Hawever they state a fixed number of 20 sentences was used. this is unclear.

Furthermore, they describe the calculation of Delta between the speech reception for SCAN and for minimic. However, they don't make any further use of this calculation.

Lastly, in the Discussion, there is need to further explain the advantage of minimic compared to the adaptive RM

Author Response

We thank the reviewer for the additional comments.

Point-by-point answers will follow in bold character.

R - although the authors have implemented important changes, there are still missing details in the methods section. They describe a relative test that finds the SNR where 50% is understood. Hawever they state a fixed number of 20 sentences was used. this is unclear.

A - for a better understanding of modern adaptive speech in noise test, in particularly the Matrix Sentence test that we used in the present paper, I suggest the reviewer to take in account the following publication:

Kollmeier B., Warzybok A., Hochmuth S., Zokoll M., Uslar V., Brand T., Wagener K.C. The multilingual matrix test: Principles, applications and comparison across languages—A review. Int. J. Audiol. 2015;54:3–16. doi: 10.3109/14992027.2015.1020971.

As the reviewer could understand, this test is meant to assess the SRT: the Speech Recognition Threshold. SRT is the lowest level of speech intensity or signal-to-noise ratio (SNR) at which an individual can accurately recognize or repeat a set of standardized speech stimuli, typically words or sentences. In this case the SRT is the minimum level of intensity to recognize the 50% of the words in 20 sentences of 5 elements. 
It is a crucial measure in audiology and speech research, providing insight into an individual's ability to understand and process speech in varying levels of background noise. The SRT serves as an important clinical parameter for evaluating the effectiveness of hearing aids, cochlear implants, and other auditory interventions, as well as assessing the impact of hearing loss on speech perception.

R- Furthermore, they describe the calculation of Delta between the speech reception for SCAN and for minimic. However, they don't make any further use of this calculation.

A- the SRT DELTA , calculated as the difference between SRTs obtained by the same subject with the use of SCAN minus the SRT obtained with the use of MiniMic2, has been further mentioned at:

• lines 154-156:  "The mean result in the SRT DELTA value [...] is 2.49 (going from –4 to 15.90)".
• lines 162-163: "When evaluating SRTs in recipients of CP1000 (M=7.77, from –1.60 to 37.80), CP950 (M=7.30, from 0.60 to 14.60) and CP900 (M=6.9, from 0.50 to 21.20) using the SCAN functionality, no statistical significance was found in their comparison. The same happened when comparing SRTs in the same recipients (CP1000, M=5.96, from –2.90 to 37.10; CP950, M=6.71, from –2.50 to 17.60; CP900, M=2.25, from –1.60 to 5.60) while making use of the MiniMic2, or  SRT DELTA values (CP1000, M=2.08, from –2.90 to 7.20; CP950, M=0.5, from –4 to 4.30; CP900, M=5.08, from 1.10 to 15.90)".
• lines 164-165: "No correlation was observed between SRTs with SCAN and MiniMic2 and the age of participants, as well as between SRT DELTA values and age".

The value of SRT DELTA has been also discussed at lines 337-341: "In detail, SRT gain reported in the study of De Ceulaer and colleagues with the source of speech at 1 meter - which is the most similar to our setting -, was 4.7dB with MiniMic 1 in comparison to SCAN, a little higher than our DELTA SCAN of 2.49dB. It is conceivable that this difference may be associated to the different test setting", making a comparison between our findings and what had been previously reported in literature.  

R- Lastly, in the Discussion, there is need to further explain the advantage of minimic compared to the adaptive RM

A- A paragraph regarding the superiority of adaptive RM has been added, according to reviewer suggestions, at lines 301-303.

Round 3

Reviewer 2 Report

accept