Dear Dr Stentiford,

Thank you very much for submitting your manuscript "Estimating orientation in Natural scenes: A Spiking Neural Network Model of the Insect Central Complex" for consideration at PLOS Computational Biology.

As with all papers reviewed by the journal, your manuscript was reviewed by members of the editorial board and by several independent reviewers. In light of the reviews (below this email), we would like to invite the resubmission of a significantly-revised version that takes into account the reviewers' comments.

Both reviewers agree on the interest of the manuscript and the scientific topics it tackles. However, they raise significant issues that we would want to see addressed before any decision can be made about acceptation. First of all, we agree with the reviews that the manuscript would greatly benefit from a systematic exploration of the major parameters (weight pattern changes with increasing revolutions, rotation angular speed distribution, size of the stimulus triangles and distance between them, spacing pattern of the R2 & R4 neurons, input strength of the PEN neurons). Moreover, a couple of modelling choices or unclear results need to be more strongly supported, including the STDP rule used for inhibitory synapses or the error increase with PEN neuron input. Finally, we also agree with reviewer 2 that the manuscript would be significantly stronger with more model validation, for instance using visual stimuli that are frequently used in fly behavioral tests. 

We cannot make any decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is also likely to be sent to reviewers for further evaluation.

When you are ready to resubmit, please upload the following:

[1] A letter containing a detailed list of your responses to the review comments and a description of the changes you have made in the manuscript. Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

[2] Two versions of the revised manuscript: one with either highlights or tracked changes denoting where the text has been changed; the other a clean version (uploaded as the manuscript file).

Important additional instructions are given below your reviewer comments.

Please prepare and submit your revised manuscript within 60 days. If you anticipate any delay, please let us know the expected resubmission date by replying to this email. Please note that revised manuscripts received after the 60-day due date may require evaluation and peer review similar to newly submitted manuscripts.

Thank you again for your submission. We hope that our editorial process has been constructive so far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Hugues Berry

Academic Editor

PLOS Computational Biology

Thomas Serre

Section Editor

PLOS Computational Biology

***********************

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #1: The paper presents a computational model focusing on orientation or heading direction within the insect central complex. While the core model shares similarities with several existing models in the literature, it incorporates more realistic visual inputs derived from recent experimental data (specifically, R2 and R4 neurons as used in this study). Tested with both simple stimuli (e.g. bright vertical bars or a linear array of triangles) and panoramic videos recorded from a moving camera in real environments, the model demonstrates its ability to estimate heading even with complex visual inputs, achieved through filtering by a small number of ring neurons with simple receptive fields, with the help of synaptic learning. While naturalistic visual inputs remain very limited, this represents a step forward. While the paper still exhibits some weaknesses, addressing them in a revision could further strengthen its contribution to the field.

Detailed comments:

The authors claim to introduce a novel spiking neural network model of the insect central complex. While the network itself may not be inherently novel given the anatomical constraints, the true innovation likely lies in a more realistic presentation of visual inputs. To better clarify the contribution of this paper, the authors should provide a more accurate description of the differences.

Regarding synaptic weight learning, the results presented are based on only a few revolutions. However, considering real animals may experience numerous rotations over a reasonable timeframe, it would be valuable to explore the mature weight patterns after learning from many revolutions.

The authors should consider systematically exploring the parameter space, including testing the effects of weight strength on the results.

Furthermore, changing the rotation angular speed or utilizing a naturalistic angular speed distribution may impact learning, particularly given the spiking-timing-dependent rule.

A closer examination of failed simulations for specific natural scenes where the activity bump was lost is warranted. Understanding the reasons behind these failures beyond correlational studies would provide valuable insights.

The effect of translational movement of the camera on the learning process appears to depend crucially on the distance from the camera to the landmarks. Therefore, documenting the distribution of distances from the camera to nearby objects is essential. Even crude estimates would prove useful for other researchers.

Fig. 2A:

some labels at the top (90 deg rotated) need more spacing

Need consistent notation: label D7 in Fig. 2 Vs delta 7 in Fig. 1

Fig 5:

B: Why are there pixel correlations at 180 deg? Bias in the video? It may cause confound with the ring activity correlation.

The organization of the panel labels is confusing. For example, “G: High correlations between ring neuron activations…” seems wrong.

Reviewer #2: The authors propose a ring-neuron network model with synthetic receptive fields with STDP (spike-timing dependent plasticity) for the insect central complex. They demonstrate that their model is able to reproduce observed neural activity and behavior under simple and natural scenes. Previously, neural network models of this brain area mostly focused on the function of the attractor dynamics and how a bump can be formed under simple stimuli. Other models showed how plastic ring circuits can form a map between visual stimuli and the ellipsoid body. This work fills the gap and provides a more complete picture of the mechanism of visual orientation of fruit flies. However, there are several critical issues the authors need to address. Here are my specific comments:

Major comments

1. The hypothesis of the hexagonal arrangement of the receptive fields of the R2 and R4 neurons seems crucial for the model. Is there any experimental evidence for the hypothesis? Otherwise, the authors should also test other arrangements, such as square lattice or random spacing. If the hexagonal arrangement delivers the best performance, this can be a testable prediction of the model.

2. The authors referred to Vogels et al 2011 for their model of plasticity for inhibitory synapses. However, unless I understand it incorrectly, the plasticity model in Vogels et al 2011 (Eqs 3-5 in Supporting Material) exhibits an opposite effect to what the authors modeled in the manuscript. Vogels et al 2011 used conductance-based synapses. Their weights represent synaptic conductance and are always positive. When postsynaptic and presynatpic spikes occur within a small time window, the weight increases, leading to strengthened inhibitory synapses. However, in this manuscript, the authors appeared to use current-based synapses, so the inhibitory weights are negative. When coincident spikes occur, the weights become less negative, or the inhibitory synapses are weakened. Can the authors elaborate? Is there any experimental support for this uncommon STDP rule?

3. It appears that the model has the best performance when the bump is solely driven by the visual input from the ring neurons and the error increases when the model takes additional input from PEN neurons (see Fig 3A, 4B, 5D & 5J). Why is that? Did the authors test whether they had a correct PEN input strength? Because the moving speed of the bump is also modulated by the PEN input, if the input strength is too high or too low, it might interfere with the ring-neuron-driven bump movement.

4. What happens if the virtual scene has a bright (white) background with one or two black stripes? This type of visual stimuli has been used widely in fruit fly behavioral tests. Flies are shown to navigate and remember the location of the stripes without problem. Does the model proposed by the authors perform equally well or is it unable to keep the bump due to the strong activation of most ring neurons? I would guess the latter, but I could be wrong. This can be a good way to validate the model.

5. It is a cool idea to test the model using triangles with different arrangements and the result is very impressive. However, not being able to discriminate between the objects (as tested in Ernst & Heisenberg 1999) is not equivalent to not being able to keep orientation (as shown in this study). It is likely that they are two completely different neural mechanisms responsible by different neural pathways. The authors need to acknowledge and discuss this in the Discussion section, or simply remove this part.

6. Following the comment above, I guess the inability to move the bump (Figure 4C) may be just a coincidence stemming from the specific sizes and arrangements of the cues and the hexagonal receptive fields. What happens if the size of the triangles and the space between them are changed? My main point is that if the authors want to keep this part, they should investigate the mechanism underlying such indiscriminability rather than just showing the phenomenon.

Minor comments

1. Line 113. Please specify the unit of tau. If the unit is millisecond, 50 ms for inhibitory synapses is way larger than typical GABAa receptors, unless the authors modeled GABAb. 100 ms for the excitatory synapses are also larger than what would be expected for cholinergic receptors. Is there any support for such large synaptic time constants? Is there any specific reason why excitatory synapses need longer time constants than that of inhibitory synapses? I can understand that longer time constants are crucial for stabilizing the dynamics of a neural circuit with a small number of neurons. The authors need to explain the reason in Methods or discuss it in Discussion.

2. Figure 2 legend, third line from bottom: missing the figure number in “(see Figure )”

3. Figure 2B. Please label the neuron indices (all or a few representative ones). Otherwise, without known which neuron is which, the connection matrices in panel C are meaningless.

4. Line 263: The phrase “median active EPG cell” is confusing. Using the word “median” makes people think that the author sorted the EPG cells by their activity level and choose the middle one.

5. Lines 323-333. Some sentences are too long and poorly written. Please rewrite them.

6. Line 332. “…resulting in multiple peaks.” Did the author mean several activity bumps in EPG cells? I do not see them in the figure.

7. Line 352. “Is” it.

8. Line 358. …in the error “measured” across all simulations….

9. Line 359. “… greater than 1% failed simulations.” Do the authors mean that they run more than 100 trials and count the number of failed trials? What do they define a "failed simulation?"

10. Figure 6. The authors may want to display a few example scenes for high, low variences and failed simulations.

11. The authors use the phrase “multi-polar ring neurons” several times in the text. However, it is confusing because “uni-polar”, “bi-polar” and “multi-polar” are usually used to describe the morphology of a neuron. The authors may want to find a better phrase for what they want to describe.

12. Line 378. “However” --> should start a new sentence.

13. Line 378. “…all groups…” What groups? High, low variances and failed simulations?

14. Line 277-382. Again, a very long sentence hard to comprehend. Please rewrite it.

**********

Have the authors made all data and (if applicable) computational code underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data and code underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data and code 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 or code —e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No: Unless I miss it, I do not see the authors mentioning any link to their code in the manuscript.

**********

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.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Chung-Chuan Lo

Figure Files:

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org.

Data Requirements:

Please note that, as a condition of publication, PLOS' data policy requires that you make available all data used to draw the conclusions outlined in your manuscript. Data must be deposited in an appropriate repository, included within the body of the manuscript, or uploaded as supporting information. This includes all numerical values that were used to generate graphs, histograms etc.. For an example in PLOS Biology see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5.

Reproducibility:

To enhance the reproducibility of your results, we recommend that you deposit your 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

Dear Dr Stentiford,

We are pleased to inform you that your manuscript 'Estimating orientation in Natural scenes: A Spiking Neural Network Model of the Insect Central Complex' has been provisionally accepted for publication in PLOS Computational Biology.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. A member of our team will be in touch with a set of requests.

Please note that your manuscript will not be scheduled for publication until you have made the required changes, so a swift response is appreciated.

IMPORTANT: The editorial review process is now complete. PLOS will only permit corrections to spelling, formatting or significant scientific errors from this point onwards. Requests for major changes, or any which affect the scientific understanding of your work, will cause delays to the publication date of your manuscript.

Should you, your institution's press office or the journal office choose to press release your paper, you will automatically be opted out of early publication. We ask that you notify us now if you or your institution is planning to press release the article. All press must be co-ordinated with PLOS.

Thank you again for supporting Open Access publishing; we are looking forward to publishing your work in PLOS Computational Biology. 

Best regards,

Hugues Berry

Section Editor

PLOS Computational Biology

Thomas Serre

Section Editor

PLOS Computational Biology

***********************************************************

Reviewer's Responses to Questions

Comments to the Authors:

Please note here if the review is uploaded as an attachment.

Reviewer #2: The authors have addressed all my comments and I would like to recommend the manuscript for publication.

**********

Have the authors made all data and (if applicable) computational code underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data and code underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data and code 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 or code —e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

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.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: Yes: Chung-Chuan Lo