Referee 1's review:

This is an interesting study that examines subjective time perception changes due to interactions between stimulus features and attention within and across the auditory and visual systems. Subjects report time dilation effects for targets that are "looming" in both systems but time compression effects in the visual domain. Interestsingly and perhaps surprisingly, there appears to be visual capture of auditory durations but not vice versa, suggesting interesting asymmetries in temporal processing of AV events, that require new theoretical models.

I think that the paper could be greatly strengthened by a more positve model rather than just a negation of existing models. Furthermore, there is very little discussion of the neural correlates underlying the observed effects.

Specific comments:

The paper could also describe the results more clearly in the following ways.

First, while describing the results, the authors only report the statistics of the F and t-tests without describing the mean effect sizes. This is important to distinguish between statistical significant effects that are meaningful from effects that are statsitically significant but negligible or small.

Second, there isn't a clear description of how the main PSE difference effects were computed. The paper says on two occassions that "cumulative Gaussian fits" and "Gaussian fits" were made to the psychometric functions. Perhaps, the authors could clarify the procedure with one example or perhaps show data for all the subjects for one of the paradigms. For instance, while the authors only report changes in the threshold, are there differences in slopes?

Third, why did the authors choose the particular 5 stimulus paradigm with the fourth one being the oddball? This doesn't appear to fall into any traditional psychophysical paradigm like 2IFC or 3IFC etc. There appears to be no justification or motivation for this paradigm.

Four, while the authors only examine stimulus duration effects, these are confounded by "perceptual intensity" cues that can be modelled by stimulus integration effects. Can they generalize their findings to intervals lapsed between unimodal and bimodal stimuli?

Five, it is a bit unclear what the authors mean by saliency outside of being an oddball. It is quite puzzling that only dilation is observed in the auditory system for "looming" but no effect is observed for the "reverse" paradigm. It is unclear if static tones as oddballs are particularly non-salient.

The data from the AV intersensory experiments should be also directly analyzed for congruent vs incongruent conditions separately before the results are compared with the forced fusion model. How much will the fact that the incongruent stimuli effects arise from asynchrony in the stimulus onset across modalities, account for the observed results?

The definitions and descriptions of the intersensory conditions are still a bit confusing.

How much could group differences contribute to differences between figures 2b and 2c? Could they have shown a "receding" effect as well? A replication of one of the conditions across group would have helped alleviate this concern.

The control for the intersensory experiments is also not clear. Also, this description of the figure as a separate auditory vs a visual effect is more confusing. Perhaps, these two sections could be integrated into one section.

Figure 2 could use the data for each condition separately as well to show the mean effects for each condition, instead of the difference with the controls.

Figure 3 could show the controls for the conditions as well.

Is there a better model that accounts for the data better?

The discussion could use a paragraph on the network of brain regions involved in AV temporal processing that are plausible substrates for observed behavioral effects.

----------
N.B. These are the general comments made by the reviewer when reviewing this paper in light of which the manuscript was revised. Specific points addressed during revision of the paper are not shown.