In this brief report, Harrison and colleagues provide a clear and precise description of a multi-site study testing relationships between anxiety and various dimensions of respiratory interoception and how this may differ in males vs. females. I think the analysis and results description were largely clear and straightforward, testing a simple and straightforward hypothesis. I have a few comments and suggestions that could help improve the paper.

1. The title could be more informative, for example, stating what the nature of the gender difference is.

2. The authors refer to 'gender', but it wasn't clear whether possible differences between biological sex and gender were measured or considered.

3. State and trait anxiety on the STAI were not clearly distinguished in my reading, either at the theoretical level or in terms of methods description. The figure indicates state anxiety. But this could be more explicit in the text, as well as why they did not examine trait anxiety. One could think this more stable aspect of anxiety might be more relevant (especially in clinical contexts) given that current states could be due to a number of contingent factors. And when were anxiety measures gathered? Was it at start of study participation, right before the task, right after the task in relation to state anxiety during performance, etc.? Further, as depression and anxiety symptoms tend to be highly correlated, and both state and trait anxiety tend to be highly correlated, can the authors rule out that results are not instead attributable to these other symptoms?

4. It could be useful to have a figure depicting the task and more information about the task experience generally. Unfamiliar readers may not understand how breathing loads feel (i.e., greater effort required during inspiration, whether this itself induces anxiety, etc.). Also, a threshold of 60-85% detection strikes me as a broad range. Why not choose something consistent like 60%? Did the authors take into account the exact detection level differences between participants in analyses? And how many trials in the staircase method were these rates typically based on to determine this percentage?

5. Lastly, while I gather this paper may have been constrained by requirements of a brief report, consideration of previous literature in the introduction and discussion struck me as quite incomplete. Especially in the discussion, there is a fairly large literature on interoception and relationships to affect and affective disorders, where many prominent studies were not mentioned that could have relevance and could be integrated with these new results. Here I have in mind work, which certainly also has limitations, by (among others) people like Desmedt, Pollatos, Khalsa, Smith, Feinstein, Paulus, Critchley (as well as other work by Garfinkel). In the one training study they mention, it is also not clarified that this is based on cardiac interoception, and it's not clear whether cardiac and respiratory interoception abilities should be related (or other modalities, such as recent work with gastrointestinal interoception). So I think a more complete consideration and integration of previous literature would be important.

I hope the authors find these suggestions for clarification useful in improving their paper.

I thought the authors responded to my comments reasonably well. I'm happy to recommend acceptance.

I did note one typo in the first paragraph that should be fixed: "...maintaining a continuously updating representation..."