PCSY-D-25-00080

Cascades and convergence: dynamic signal flow in a synapse-level brain network

PLOS Complex Systems

Dear Dr. Betzel,

Thank you for submitting your manuscript to PLOS Complex Systems. After careful consideration, we feel that it has merit but does not fully meet PLOS Complex Systems's publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript within 60 days Dec 12 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at complexsystems@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pcsy/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

* A rebuttal letter that responds to each point raised by the editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. This file does not need to include responses to any formatting updates and technical items listed in the 'Journal Requirements' section below.

* A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

* An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, competing interests statement, or data availability statement, please make these updates within the submission form at the time of resubmission. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

We look forward to receiving your revised manuscript.

Kind regards,

Y-h. Taguchi, Dr. Sci.

Academic Editor

PLOS Complex Systems

Y-h. Taguchi

Academic Editor

PLOS Complex Systems

Hocine Cherifi

Editor-in-Chief

PLOS Complex Systems

Journal Requirements:

If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise.

Additional Editor Comments (if provided):

Since the reviewer 3 seems to have serious concerns about the relationship with real biology, the author should address this point comprehensively.

[Note: HTML markup is below. Please do not edit.]

Reviewers' Comments:

Reviewer's Responses to Questions

Comments to the Author

1. Does this manuscript meet PLOS Complex Systems’s publication criteria? Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe methodologically and ethically rigorous research with conclusions that are appropriately drawn based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available (please refer to the Data Availability Statement at the start of the manuscript PDF file)?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception. The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS Complex Systems does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Betzel and colleagues run various forms of cascade models on the adult Drosophila connectome. Using this framework, they demonstrate how the propagation of activity from different sensory modalities spreads across the network, and show how these spreading patterns from different modalities converge and interfere with one another. These spreading patterns are aligned to basic features of the structural connectome, giving insight into how structural properties shape network dynamics.

This is a nicely done paper, I enjoyed reading it. The model is elegant and insightful. I think all my comments can be considered minor, there are just many of them!

1. Why do the visual modalities show such a distinctive, delayed trajectory? It seems interesting. For the optic lobe (is this also compound?), it appears to contain the most neurons. Is this related to their sparsity at all (a cascade has trouble escaping the visual networks) or is there another possible reason? Do the authors have any insight?

2. If you ran the same cascade model on an undirected network, would you expect radically different results? I don’t think you need to redo all the analysis with an undirected case, but a quick comment wouldn’t hurt.

3. The discussion makes little reference to fly neural physiology/biology. There is this mention “This proximity between sensors and effectors is consistent with behavioral observations and suggests streamlined sensorimotor pathways.” But I think this could be unpacked more (explain how this pathway could produce a specific behaviour) and some references to these prior observations would be handy! In general, the paper would be strengthened by some more concrete links between the model cascading dynamics and observed neural dynamics and/or fly behaviour.

4. Regarding Figure 1, it is a nice illustration, but I have some small suggestions. First, the ellipses for panels B and C appear to have migrated to the centre of the figure so you might want to move them back to the right hand side. Second, the cascade is progressing along the underlying network, but this is not shown on the first timepoint. I think you can just show the faded lines indicating connections across all the plots, with them in black to indicate active connections (as you have already done). Third, I would also add in the caption something brief about how competition for activating a given node from different sources is resolved.

5. In the introduction you say you ran three different cascade models, but by my count there is four? Because you have one model where you are selecting seeds randomly across all modalities, which seems distinct from the unimodal, cooperative, and competitive models presented in the introduction. I would clarify, and perhaps adjust Figure 2 to add this additional model.

6. When first mentioning the connectome at the start of the methods, indicate that it is directed and weighted.

7. In Figure 2, for panel c, say “Note that in c the color and labels for density are on a logarithmic scale” because I was thrown at first by what you meant by labels being on a scale! By others, does this mean all non-sensory neurons? In 2c, the dot for thermo-thermo connectivity is not centred. Also in Figure 2, maybe add a spatial map showing the positions off the different neuron system types (and can also include all the other non-sensory neurons)?

8. “Optic lobe” is listed as a sensory modality, but elsewhere in the paper it appeared to be described as “compound”? Is this one and the same? If so (and if not so!) this should be clarified.

9. I would also add in somewhere at the start that all the interesting stuff happens in the first steps of the model (can refer to Figure S1) and maybe give some indication of when the model reaches its end state. Additionally, when changing the transmission probability for a fixed seed, I assume you’ll get roughly the same spatial pattern of activation, it will just occur over a different timeframe?

10. In the section “Edge contributions to cascades”, you mention calculating the number of active synapses in the model. However just prior to this when discussing the model, it is only discussed in terms of neurons being activated. While it is briefly mentioned that you can also measure activity in individual synapses, I would add in something here so people understand how you determine a synapse as being activated (I assume a synapse becomes active if the neuron it is attached to becomes active?).

11. Relating to the above point, you calculate the total, unique and ration of synapses activated across simulations. This wasn’t clear to me on my first read through (the first mention that different simulations are run to get these measures comes after you describe these measures). I suggest the following: “To address these questions, we ran multiple simulations (N = 1000) where Nseed = 16 sensory neurons were randomly drawn across all modalities and activated with ptransmission = 0.01. At each time step, we calculated across all simulations three quantities: the total number of active synapses, the number of unique synapses used, and their ratio (unique/total).”

12. Figure 3. Panel d, I seem to have missed what all the abbreviations in the legend stand for? Panel e. For the sub-class matrices, the timepoint label is partially obscuring the matrix. Also in e, some of the lines marking out the rows aren’t displayed, I am assuming this is just a rendering error caused by whatever pdf conversion is done and should be fixed in the final version? Double check just in case. Also a colourbar would be welcome for usage counts!

13. “dynamic and structural centrality” are not explicitly defined on page 4. Suggest the following tweak “we computed the outdegree of each neuron based on edge usage (dynamic centrality) and compared this to their degree and strength in the underlying connectome (structural centrality)”

14. I had difficulty understanding Figure 4h at first, because I assumed the numbers corresponded to the number in that section of the pie chart, rather than it being an index of which regions make up a given combination. This can be easily fixed by adding commas between the numbers.

15. For the cooperative cascade, it is stated one of the conditions is a comparison to a model where “all neurons in all sensory modalities are seeded”. Does this mean that a) a subset of all sensory neurons could be selected as a seed for each model run (and how does this differ from the very first model) or b) does it really mean that every single sensory neuron was selected as a seed.

16. For the unimodal, competitive and cooperative cascades, I assume a subset of neurons from each modality are selected as seeds across each model run? I wouldn’t mind a clarification of this, as well as how many neurons were selected as seeds each time (can just put this in the methods e.g., “For all unimodal, cooperative, and competitive cascades, 16 neurons from each sensory modality being examined were randomly selected as seeds”)

17. For Figure 6a, shouldn’t the arrows point to the difference map? Because the difference map is created from the input of the unimodal and competitive probabilities, whereas to me the arrows pointing out suggest the different map is input to the other two maps.

18. In calculating the differences in activation probability for the cooperative model, when this is first mentioned on page 9 perhaps explicitly make clear this is calculated within a given sensory modality e.g., “We highlight these distinctions by calculating the difference between the competitive and unimodal activation probability for each sensory modality”.

19. Figure 8a, the text appears to imply that the clustering was done on this matrix, but the methods say it was done on a different data matrix? As 8c appears to drop a number of columns from 8a, if so perhaps also restrict 8a to only showing the first X timesteps (also need a label someone describing what the columns are)?

20. Figure 5d, 7f, 8d appear to be missing some column labels

Stuart Oldham, Murdoch Children’s Research Institute

Reviewer #2: Betzel, et al. utilise a cascade spreading model to bridge nanoscale connectomics with system-level network neuroscience. While most network theory has focused on meso- and macro-scale graphs, the authors demonstrate that using similar tools and principles, it is possible to gain insight into the interaction between anatomy and cascade behaviour at synapse-level resolution.

Overall, I congratulate the authors on technically sound work with a very interesting dataset. It was well written, and I believe it will be of interest to several intersecting fields.

I have no major concerns but offer a list of minor comments and suggestions. I wrote the comments approximately in order of reading the paper, though because there are no line numbers in the paper, I couldn't reference precise locations.

Minor concerns/comments/suggestions:

Introduction

• The authors' use of a spreading model to explore how signals propagate from sensory systems in Drosophila, from the segregated periphery to the convergent shared higher-order structures, is interesting. The approach of unimodal (single seeding of a modality), cooperative (joint influence), and competitive (competition for influence) scenarios is novel – but how is this biological? This seems more of a general description rather than having any theoretical conclusion at the end of the introduction. What are the neurons and connections facilitating early, rapid activation? It would be good to make these lessons clear to the reader. I would suggest summarising the core results at the end of the introduction.

• For Figure 1, it would be helpful to clarify the features of this model, which could be seen as advantageous or disadvantageous, particularly a discrete time spreading model (as described: https://pubs.aip.org/aip/cha/article-abstract/34/4/041501/3282989/An-integrative-dynamical-perspective-for-graph?redirectedFrom=fulltext) that does not consider synaptic/dendritic delays, which would lead to less synchronous updates to outgoing nodes (depending on the scale, this could also be axonal). Of course, these assumptions may be totally fine! But it would be good to explain why they are appropriate. This would potentially also be interesting to speculate about in the discussion.

Results

• Excellent description of the quite considerable sparsity of the sensory subgraph, including its density relative to a random sub-sample, and the modularity. This is very interesting! It would be helpful to understand more about the hierarchy of modularity here: can you report the modularity within each annotation? For example, it may be that some sub-sensory systems are themselves more parallel than others, as a function of the modality.

• "Taken together, these results indicate that direct sensory-sensory interactions are both rare and strongly modality-specific. This organisation suggests that multimodal integration likely occurs through polysynaptic pathways that recruit non-sensory populations." Can you make this final sentence clearer? It is totally correct – given the parallelism/modularity of the sensory neuron (direct) matrix, for there to be integration it must happen outside of these neurons (i.e., non-sensory populations) which definitionally is more than one synapse (i.e., polysynaptic) – but the wording is not very clear. This also requires more motivation because it is used as the justification to introduce the cascade model, which is important. (Could you perhaps avoid using "recruit"? This seems more of a dynamic term, rather than describing synaptic connectivity. Though I understand you are bringing in a dynamic model on top of the graph).

• "At each subsequent time step, active neurons attempt to activate each of their postsynaptic partners with probability p_transmission." – I presume this probability comes from the normalised weight. If so, please state this explicitly; if not, please describe the derivation (actually, coming back to this, I realise it is a single value – but still justify this, as the cascade model just utilises the binary matrix). In general, the model needs more description of how it works, e.g., what is the refractory state, and how do you justify any parameter values you use? I know it is in S1, but more detail is needed here, as it is important for reading the rest.

• Why N_seed=16? Why p_transmission = 0.01? I imagine you have good reasons, but they're not explained. I'd also like to see your results outside of this parameter range.

• Figures 3a and 3b are not described in the text. Please describe them in order.

• Is "into the rest of the brain" justified? The terminal location of these cascades is not described, as far as I can tell. In principle, the cascade could be oscillating between sensory and non-sensory regions back and forth, though this seems unlikely.

• In Figures 3d,e, can you integrate other properties that could be predictive: such as centrality or communicability? Cell-type (or other features one can uniquely get from this dataset)? This seems like a critical part of the story. I gather it is very distributed, and that position in the hierarchy matters much more than cell type. But it would be valuable to confirm this or identify any interactions between cell type and graphical properties.

• At Figure 4c, you concatenate probability vectors: I think it would be helpful here or previously to better describe what these probabilities represent. Previously there was a scalar p_transmission, but now I believe this is derived from the network's weights.

• Is the "convergence zone index" not a slightly confusing term, given it is defined solely by activity level of a unimodal input? That is, it is convergent only insofar as it exceeds an arbitrary threshold. Please justify the 25% threshold. I believe you are using the term convergence because you're suggesting it is now a cascade that has globally spread (and hence, converged with the rest of the brain?). I may be missing something. [Coming back to this, I now see you outline this at the beginning of the cooperative sensory cascades section. I still believe using "convergence" previously is confusing, because it is not clear what is converging at all (I agree that a more realistic extension that allows cascades to interact would constitute a convergence).]

• In this section, please describe the termination locations of the cascades, broken down by input seed. This would be of general interest, particularly when considering future work and implications for behaviour and behavioural models that deal with multi-sensory integration.

• The cooperative and competitive cascade scenarios are very interesting. But you need to clearly spell out how you manipulate the model to promote cooperative versus competitive cascades.

• In the cooperative case, it seems quite straightforward: you stimulate simultaneously different modalities of sensory neurons. But it took me some time to understand what you meant in the competitive case – "wherein the signal propagated from two sensory modalities retains modal specificity. That is, neurons activated by either of these two modalities inherit the 'label' of their respective sensory activator, maintaining the distinct identity of each signal."

• I think what you do therefore is run the same simulation but label the updated neurons in terms of the original signal. Nothing is different, just the labelling. If so, please make this explicitly clearer because it is more of an interpretation of how a simulation runs, rather than the simulation itself. Also, please outline the case in which one signal would indeed dominate at the convergence point – that is, when two cascades meet a neuron, what dictates it becoming dominated by one over another?

Discussion

• It would be interesting if you discussed how this interacts with the wave literature on interfering waves in cortex, which show constructive and destructive behaviours. Of course, we have different spatial scales here, so I wouldn't necessarily expect waves, but it seems to resonate with similar principles. It would be valuable for you to explain this connection.

• You reference decision-making, planning, etc. in the link to multimodal integration. I would suggest speculating on models which concern multisensory integration can be integrated concretely with your types of dynamical models that act on graphs (e.g., such as yours, or outlined e.g., https://pubs.aip.org/aip/cha/article-abstract/34/4/041501/3282989/An-integrative-dynamical-perspective-for-graph?redirectedFrom=fulltext). For example, this paper may be of interest: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1012246 This would be very interesting and would more clearly outline how we can move from anatomy to cascades to behaviour.

Reviewer #3: This manuscript presents a computational study of sensory signal propagation in the adult Drosophila connectome using a network-based spreading model. The topic is interesting and relevant, and the idea of combining structural connectivity with dynamic simulations is promising.

However, several aspects need clarification and improvement before publication:

1. Model justification: The spreading model and its cooperative/competitive dynamics are conceptually described, but the biological interpretation of the parameters and mechanisms should be explained more clearly.

2. Validation: It would strengthen the paper to compare the simulation results to experimental data or provide qualitative evidence supporting the predicted convergence zones.

3. Robustness: A sensitivity analysis or discussion of parameter effects would help show whether the findings are stable.

4. Interpretation: The functional significance of convergence zones and neuron classifications could be discussed in more detail, including testable hypotheses.

Overall, the study is promising, but major revisions are needed to clarify the model, support the conclusions, and improve the biological interpretation.

Recommendation: Major Revision

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

Do you want your identity to be public for this peer review? If you choose “no”, your identity will remain anonymous but your review may still be made public.

For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Stuart Oldham

Reviewer #2: No

Reviewer #3: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

Figure resubmission:

While revising your submission, we strongly recommend that you use PLOS’s NAAS tool (https://ngplosjournals.pagemajik.ai/artanalysis) to test your figure files. NAAS can convert your figure files to the TIFF file type and meet basic requirements (such as print size, resolution), or provide you with a report on issues that do not meet our requirements and that NAAS cannot fix.

After uploading your figures to PLOS’s NAAS tool - https://ngplosjournals.pagemajik.ai/artanalysis, NAAS will process the files provided and display the results in the "Uploaded Files" section of the page as the processing is complete. If the uploaded figures meet our requirements (or NAAS is able to fix the files to meet our requirements), the figure will be marked as "fixed" above. If NAAS is unable to fix the files, a red "failed" label will appear above. When NAAS has confirmed that the figure files meet our requirements, please download the file via the download option, and include these NAAS processed figure files when submitting your revised manuscript.

Reproducibility:

To enhance the reproducibility of your results, we recommend that authors of applicable studies deposit laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Cascades and convergence: dynamic signal flow in a synapse-level brain network

PCSY-D-25-00080R1

Dear Dr. Betzel,

We're pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you'll receive an e-mail detailing the required amendments. When these have been addressed, you'll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at https://www.editorialmanager.com/pcsy/ click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. For questions related to billing, please contact billing support at https://plos.my.site.com/s/.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they'll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact complexsystems@plos.org.

Kind regards,

Y-h. Taguchi, Dr. Sci.

Academic Editor

PLOS Complex Systems

Additional Editor Comments (optional):

The authors addressed the reviewer's comments well.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

--------------------

2. Does this manuscript meet PLOS Complex Systems's publication criteria? Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe methodologically and ethically rigorous research with conclusions that are appropriately drawn based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

--------------------

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

--------------------

4. Have the authors made all data underlying the findings in their manuscript fully available (please refer to the Data Availability Statement at the start of the manuscript PDF file)?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception. The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

--------------------

5. Is the manuscript presented in an intelligible fashion and written in standard English?<br/><br/>PLOS Complex Systems does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

--------------------

6. Review Comments to the Author<br/><br/>Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I feel all my comments have been more than adequately addressed :) I appreciated the effort the authors have gone to in addressing my points and I have no issue recommending for publication. I look forward to seeing the full paper in print!

Stuart Oldham, Murdoch Children’s Research Institute

Reviewer #2: Thank you for the thorough response, addressing all my comments & congratulations on the publication.

--------------------

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

Do you want your identity to be public for this peer review? If you choose “no”, your identity will remain anonymous but your review may still be made public.

For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Stuart Oldham

Reviewer #2: Yes: Danyal Akarca

--------------------