The authors investigated if increased levels of nitric oxide (NO) in the ventral cochlear nucleus (VCN) contributed to the increased gain applied to input from the auditory nerve after the induction of tinnitus caused by unilateral noise exposure in guinea pigs. They discuss their results in the context of the theory of homeostatic plasticity (HP) and find that, despite permanent threshold changes in all noise-exposed animals, the driven firing rates of neurons in the VCN of tinnitus (T) animals are maintained at similar levels to those seen in controls while the mean driven firing rate of non-tinnitus (NT) animals is significantly reduced. Blocking endogenous NO led to a decrease of mean firing rate mainly in neurons of T animals but hardly in neurons of NT animals indicating that NO is facilitating the mean firing rate in T animals.

In general, I think this paper has a high value in understanding the neuronal substrate of tinnitus development. The language is very good. The methodology is well explained, the neurophysiology is sound and the behavioral experiments are well done. Nevertheless, I think some more analysis in the context of the behavior will improve the paper significantly. Additionally, in the discussion a somewhat broader view beyond the context of HP may widen the impact of the manuscript also.
I suggest a major revision, after which I would be happy to accept the paper.

Major Points
Methods:
The analysis of the gap prepulse inhibition of acoustic startle (GPIAS) is tricky. The test of pre and post exposure amplitudes has to be reanalyzed carefully, as is has been shown, that the amplitudes are log-normal distributed. In the paper of Schilling et al., 2017 (https://doi.org/10.3389/fnbeh.2017.00198) a very good analysis method for this kind of data has been proposed. Finally, you receive an effect size for every single stimulus tested and you can even perform t-tests on the log-normalized PPI distributions of pre and post exposure data. On that level, one can interpret negative effect sizes with a significant t-test as tinnitus frequencies. With the effect sizes you can also perform regression analyses for the neurons responding to corresponding frequencies and therefor (maybe) refine your arguments, that only a part of the neurons change their responses during NO-manipulation. Maybe these are the neurons that lie within a specific relationship to the tinnitus frequency.
Results:
With this GPIAS analysis you can also show the effect sizes and/or number of tinnitus frequencies for all animals (also NT) in Fig 3. That would be more informative compared to the exemplary data given now.
You can also calculate regression analyses for the T and NT neurons of the animals as suggested above. This can be integrated into Figs. 5 and 7.
Discussion:
Page 27 – Potential role of NO…: Here you should broaden the view of the discussion beyond the HP theory. For example the Stochastics resonance hypothesis of Krauss et al, 2016 and 2017 (https://doi.org/10.3389/fnins.2016.00597; https://doi.org/10.3389/fneur.2017.00031) could also explain (at least partially) the interaction of VCN, DCN and input from the auditory nerve. In this context, also the papers from Liberman about synaptopathy and neuropathy in the cochlea/auditory nerve (e.g. https://doi.org/10.3389/fnsys.2018.00059 and others) should be discussed.
Minor points:
Results:
Page 15: define s.d. at the beginning of the paragraph and use ± instead: e.g. 27.4 (± 6.5) dB
Page 18/Fig 5: Change the order of the subpanels or the order of appearance in the manuscript. You start with Fig 5H and then jump to Fig 5A. The order of the other subpanels seems then OK.