Appropriate data segmentation improves speech encoding models: Analysis and simulation of electrophysiological recordings

Ole Bialas; Edmund C. Lalor

doi:10.1371/journal.pone.0323276

Peer Review History

Original SubmissionSeptember 21, 2024
21 Sep 2024 Author Response https://doi.org/10.1371/journal.pone.0323276.r001
8 Oct 2024 Decision Letter - Li Yang, Editor PONE-D-24-38627Appropriate data segmentation improves speech encoding modelsPLOS ONE Dear Dr. Bialas, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’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. ============================== Editor Comments:Thanks for submitting your work to PLOS ONE. Your manuscript has now been assessed by our editorial team and external peer experts. While they found it interesting, you will see that they have raised many serious problems and are advising that you should revise your manuscript thoroughly. At the same time, please submit the point-by-point responses to reviewers' comments. If you are prepared to undertake the work required, I would be pleased to reconsider my decision. Please note that this revision decision does not assure the acceptance of your work. Thanks for the opportunity to consider your work. Editor Reviews:1.Please further re-evaluate and clarify the rationality of the methodology, which is linked to the result reliability and scientific rigor.2.Please consider to comprehensively propose the strengths and limitations of your work.3. Please consider to use structured abstract.4. I did not see conclusion section. Please add this part.5. Please consider to add some tables to summarize the core data of your analysis and results. I think it will be clearer for readers to grasp your findings.6. Minor suggestion: improve the quality of your figures (both in resolution and contents).============================== Please submit your revised manuscript by Nov 22 2024 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 plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ 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 academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. 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, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Li Yang, M.D. Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. Please update your submission to use the PLOS LaTeX template. The template and more information on our requirements for LaTeX submissions can be found at http://journals.plos.org/plosone/s/latex. 3. Please include your full ethics statement in the ‘Methods’ section of your manuscript file. In your statement, please include the full name of the IRB or ethics committee who approved or waived your study, as well as whether or not you obtained informed written or verbal consent. If consent was waived for your study, please include this information in your statement as well. Additional Editor Comments: Please address the concerns proposed by Editor Reviews. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Partly Reviewer #2: Partly ******** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes Reviewer #2: I Don't Know ****** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data 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 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: No ****** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE 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 ****** 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: Thank you for the opportunity to review this manuscript, which presents an important exploration of EEG responses to speech segmentation. I appreciate the authors’ efforts in conducting this research and believe that with some revisions, the paper has the potential to make a valuable contribution to the field. However, it is important that clear rationales for their methodological decisions are provided, as this will enhance the transparency and reproducibility of the findings. Detailed explanations of participant selection criteria, the choice of EEG data, and the specific segment lengths used would strengthen the overall clarity and impact of the manuscript. Additionally, the figures are almost publication-ready; they just need a few clarifying details. I have attached a document with more detailed comments for your consideration. I look forward to reading future installments of this study. Reviewer #2: In this paper the authors have used simulations and EEG recordings to investigate how non-stationarities affect temporal response functions (TRF) models for continuous speech processing. Based on the findings it is suggested that non-stationarities may reduce the performance of TRF models. However, it can be partially compensated by dividing the data into shorter segments that approximate stationary. This is a topic of interest. For authors of this paper, I have the following comments. 1. The expression description of the paper needs improvement. The introduction section should be improved by clarifying the similarities and differences between the related work and the proposed method. In current form the contribution of the paper seems marginal. Authors should emphasize and clearly describe their contribution, preferably in the form of bullets. 2. Usually, a system used for evaluation purpose has a very careful design about the front-end sensors. However, no physical implementation and parameter detail is provided in this work. 3. The used materials and methods should be clearly described and clearly specify the implementation details/parameters of all methods. 4. Provide a URL of the studied dataset or share the studied dataset in csv format for an ease of review process. 5. Properly support claims in the discussion section by providing quantitative findings. Also present the performance comparison with counterparts, preferably in a tabular form while focusing on the main methods and major findings. ****** 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. 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: 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.] 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. Registration is free. 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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. Attachments Attachment Submitted filename: PONE-D-24-38627_comments.pdf https://doi.org/10.1371/journal.pone.0323276.r002
Revision 1
12 Jan 2025 Author Response Dear Editor and Reviewers, Thank you for providing us with the opportunity to refine and clarify our manuscript. In the following, we will address the points you have raised. Where it is appropriate, we will refer to edited lines in the revised manuscripts (edits are marked red). Comments by the Editor 1.Please further re-evaluate and clarify the rationality of the methodology, which is linked to the result reliability and scientific rigor. A: We focus on TRF models because they are the most commonly used kind of encoding model and they are comparatively small and simple (see second paragraph of the introduction). We analyze simulated data because they allow us to test our hypothesis under a known ground truth and we use EEG recordings to test our hypothesis on real data (see ”Methods” part of the abstract and the first paragraph in the ”Methods” section). We chose this data set because it has proven to be well suited for TRF analyses and the individual recordings are long enough to test our models across a wide range of segment durations (see section ”Modeling EEG responses to speech”). 2.Please consider to comprehensively propose the strengths and limitations of your work. A: The strength of our approach lies in the combination of simulations and EEG analysis. This way we can test our hypothesis theoretically, under a known ground truth and empirically, on actual neural data. We added these considerations to the abstract as well as the discussion section (paragraph 4). Our results show that segmentation is an easy method for improving the accuracy of speech encoding models. This is relevant because these models are widely used in the neuroscience community, potentially including medical diagnostics. 3. Please consider to use structured abstract. A: We changed the abstract to use a structured format. 4. I did not see conclusion section. Please add this part. A: We have added a Conclusion section. 5. Please consider to add some tables to summarize the core data of your analysis and results. I think it will be clearer for readers to grasp your findings. A: We could add table that contains the average model accuracies for each model and all segment durations. However, we think that the absolute model accuracy is not relevant here, since the key aspect is the relationship between model accuracy which is depicted in Fig. 3. 6. Minor suggestion: improve the quality of your figures (both in resolution and contents). A: We re-created the figures with higher resolution. Comments by Reviewer I 30-31: While read speech may approximate “real speech,” it still lacks many features of truly spontaneous speech. To substantiate this claim, consider citing the following studies: Howell, P., & Kadi-Hanifi (1991). Comparison of prosodic properties betweewn read and sponta- neous speech material. Speech Communication, 10(2), 163-169. https://doi.org/10.1016/0167- 6393(91)90039-V Mehta, G., & Cutler, A. (1988). Detection of target phonemes in sponta- neous and read speech. Language and Speech, 31( Pt 2), 135-156. https://doi.org/10.1177/0023830988031002 A: Many thanks to the reviewer for pointing us to this interesting literature. We fully agree that read speech may deviate from real conversational speech in important ways. However, in the case of our study, we are not making any argument about how speech is processed in the brain. Rather, we are investigating how the encoding models typically used in the speech literature are affected by the statistical properties of the neural recordings. Nonetheless, we have added a small caveat to the discussion on the nature of the speech materials we used. 47-49: Please clarify the definition of “estimand” to assist readers who may be unfamiliar with the term. A: Rather than providing a definition, we added an example to clarify that predictor refers to speech features and estimand refers to the neural recording 121-122: Upon reviewing Di Liberto et al. (2015) and Broderick et al. (2018), neither article specifies the actual material read by the speaker. If the original text is unavailable, could details regarding the narrative style or genre be provided? Understanding the type of text used could impact the interpretation of results. A: The story is ”The Old Man and the Sea”. We added this information to the subsection ”Modeling EEG responses to speech”. What criteria were used for selecting the 19 participants in this study? A: The subjects were all healthy neurotypical adults, we added this information to the subsection ”Modeling EEG responses to speech”. There were no other special criteria applied. The data were collected for Di LIberto et al., 2015 and Broderick et al., 2018 and are described in more detail in those studies. Why was this particular EEG database chosen for the analysis? A: We chose this data set because it has proven to be well suited for TRF analyses and the individual recordings are long enough to test our models across a wide range of segment durations. 131–136: The inclusion of the number of spectral bands as a variable is a valuable addition; however, they are infrequently referenced in the results and discussion sections. It would be beneficial to discuss their impact on the findings. A: We found that the accuracy improvement from segmentation was larger for the 16-band model compared to the 8- and 1-band models. This could be because the increased number of parameters increases the extent to which the model (over-)fits the non-stationary trends in the data. We added this to our discussion section. 143-145: Please clarify the source of the “50 minutes” of data, as the analysis seems to involve only 19 subjects, each contributing between 170 and 200 seconds of data. A: Thank you for pointing out the ambiguity here. There are 50 minutes of recordings for every subject, we rephrased the paragraph to make this clearer. 143-145: What was the rationale for choosing these specific segment lengths? Given that Gonen and Tcheslavski (2012) identified a critical duration of 2 to 25 seconds, why not test segments up to 25 seconds using the current dataset? A: The durations were chosen so that the 50 minutes of data could be divided evenly into segments for every duration (we added this expla- 2 nation to the manuscript). The studies that identified critical durations between 2 and 25 seconds tested the randomness or gaussianity of the recording. Because our approach was different we chose to include the widest possible range of values to minimize a priori assumptions about the relevant timescale. However, our results seem to agree with the mentioned studies since there is little difference between using segments that are 20 seconds and longer. We added a sentence to both the Methods and Discussion sections to relate our segment durations to the previous studies 145-146: An exclusion criterion of 0.01 for model accuracy appears unusually low. Please provide justification for this threshold. A: It is important to note that we are attempting to model unaveraged EEG - most of which is unrelated to the speech stimulus. It is quite typical for such models to produce prediction accuracy scores of r ¡ 0.1. The point of the exclusion criterion was to reject participants where the model could not explain any variance in the data. For the one subject that was rejected due to the criterion, accuracy approximated 0 for all models (average r was 0.001). We added this latter information to the paragraph 191-193: “On average” is stated twice in this sentence. A: We corrected the sentence 198: It might be beyond the scope of this paper, but was any post hoc analysis conducted to determine the factors driving variability across subjects? A: This was to motivation for the linear regression that relates the model improvement from segmentation to the baseline prediction accuracy. The fact that the improvement was larger for subjects where the baseline accuracy was low suggests that the non-stationarity is not related to stimulus driven activity. However, since interindividual variability in EEG recordings is another area of research, we do consider further investigations into this beyond the scope of this paper indeed. 202-204: The phrase, “What’s more, there was a negative relationship (linear regression, r = -0.62, p = 0.006) between the relative change in accuracy due to segmentation. . . ” needs clarification. Specifically, which segmentation is referenced here: the average of 10 and 120 seconds, the average across all segment durations, or the optimal segmentation for each sub- ject? A: This is indeed ambiguously phrased. For every subject we took the relative difference in accuracy between 10 and 120 second segments (i.e. the improvement due to segmentation) and the peak prediction accuracy (i.e. how well the subjects EEG could be predicted with the best combination of parameters). These two estimates were negatively correlated across subjects. We rewrote the paragraph to make it clearer. 206–208: Is there any commentary regarding the identified outlier? A: It’s possible that the recordings of the outlier were unusually stationary. As our simulations showed, segmentation does not improve but instead reduces model accuracy when the data are stationary. We added this commentary to the paragraph. 209: Consider using “improve” or “increase” instead of “affect” to more powerfully convey the direction of the change. A: We changed the wording of the sentence. Figures Although it is common in the field, including a definition for “a.u.” would be helpful for readers. A: We added an explanation to the caption of each figure. 3 Please correct the x-axis label from “segmend” to “segment.” A: We fixed the typo. Figure 3a & c: Although both y-axes are represented in a.u., are they on the same scale? For instance, does a 0.1 change in accuracy in panel c correspond to a 0.1 change in accuracy in panel a? If so, it would be helpful to specify this in the figure legend. A: No, they are not the same. Figure 3a shows accuracy normalized to each model’s maximum and 3c shows the relative difference between 10s and 120s segments. We modified the caption to make this more explicit. Figure 3c: Regarding the “Change in accuracy,” which segmentation level was used to ob- tain these values? A: This is the relative difference between 10s and 120s segments. Because this information would have been too long to put it in the axis title, we put it in the figure caption While the Ethics statement is appreciated, it is somewhat confusing as neither author is affiliated with Trinity College Dublin. It would be helpful to include a sentence clarifying that the EEG data originated from Di Liberto et al. (2015) or Broderick et al. (2018) at that institution, allowing readers to understand the context without needing to consult those studies. A: Personally, we think that an ethics section is not strictly necessary since we only used publicly available data. However, this seems to be a requirement by the journal. Your suggestion is a good way of making the relationship between our study and the data more explicit. Comments by Reviewer2 1. The expression description of the paper needs improvement. The introduction section should be improved by clarifying the similarities and differences between the related work and the proposed method. In current form the contribution of the paper seems marginal. Authors should emphasize and clearly describe their contribution, preferably in the form of bullets. A: There are two ways in which the paper contributes to the field: 1. We probe the non- stationarity of EEG recordings by quantifying the improvement of model accuracy due to data segmentation and by using simulations to show that these improvements can be explained by non- stationary trends in the data 2. We show that the common practice of fitting speech encoding models to long, continuous recordings is suboptimal and urge researchers to adopt segmentation as part of their model fitting process. Regarding the differences between ours and related work, see the last paragraph of the subsection ”Are neural recordings stationary?”. We also hope that the added structured abstract and conclusion sections add some clarity with respect of the points you raised. 2. Usually, a system used for evaluation purpose has a very careful design about the front-end sensors. However, no physical implementation and parameter detail is provided in this work. A: The data we present in this manuscript came from a publicly available EEG dataset. We note that the data are described in two previously published papers (Di Liberto et al., 2015 Broderick et al., 2018). Those papers describe the data collection hardware in detail. 3. The used materials and methods should be clearly described and clearly specify the imple- mentation details/parameters of all methods. A: We have attempted to be clear about the simulation and EEG analysis parameters throughout the methods and results sections. Also, as mentioned be- low, we have made all of our code available so that others can see precisely what parameters we used for each step. 4 4. Provide a URL of the studied dataset or share the studied dataset in csv format for an ease of review process. A: The data is hosted on OpenNeuro and can be obtained, together with all of our code, from a GitHub repository. Both, repositories are linked in the ”Data availability” subsection in ”Materials and methods”. The data are shared in the standardized BIDS format — csv would be a very inefficient way of sharing a large EEG data set. 5. Properly support claims in the discussion section by providing quantitative findings. Also present the performance comparison with counterparts, preferably in a tabular form while focusing on the main methods and major findings. A: We have made extensive revisions to the results and discussion sections and hope that both are now coherent in terms of the quantitative findings and the discussion of same. Sincerely, Ole Bialas and Edmund Lalor Attachments Attachment Submitted filename: rebuttal.pdf https://doi.org/10.1371/journal.pone.0323276.r003
3 Feb 2025 Decision Letter - Li Yang, Editor PONE-D-24-38627R1Appropriate data segmentation improves speech encoding modelsPLOS ONE Dear Dr. Bialas, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’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 by Mar 20 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 plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ 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 academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. 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, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Li Yang, M.D. Academic Editor PLOS ONE Journal Requirements: Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice. [Note: HTML markup is below. Please do not edit.] 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 ******** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ****** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: N/A ****** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data 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 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 ****** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE 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 ****** 6. 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: Thank you for the thorough and thoughtful revisions. I appreciated the explicit outlining of how this study is distinct and contributes to the field, as well as the clear statement of the hypothesis. The explanation for the low data exclusion threshold and discussion of the outlier were also helpful. The concluding paragraph is a welcome addition and effectively summarized the study. Additionally, I appreciated the clarification regarding the incompatibility for comparing Figure 3 and a.u.s. Thank you for specifying that the reading passage was “The Old Man and the Sea.” For unfamiliar readers, could the authorship or a brief description of the story be added? This is particularly relevant as different emotional states may yield different EEG results. Good luck with this line of research! ****** 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. 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 ******** [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.] 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. Registration is free. 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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0323276.r004
Revision 2
12 Mar 2025 Author Response Dear Reviewer 1, Thank you for your comments. In your response, you requested: "For unfamiliar readers, could the authorship or a brief description of the story be added?" We have added two sentences to the section "Modeling EEG responses to speech" that describe the content and tone of the narration to allow the reader to assess the emotional impact. Sincerely, Ole Bialas and Edmund C. Lalor Attachments Attachment Submitted filename: rebuttal2.pdf https://doi.org/10.1371/journal.pone.0323276.r005
18 Mar 2025 Decision Letter - Li Yang, Editor PONE-D-24-38627R2Appropriate data segmentation improves speech encoding modelsPLOS ONE Dear Dr. Bialas, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’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 by May 02 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 plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ 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 academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. 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, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Li Yang, M.D. Academic Editor PLOS ONE Journal Requirements: Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice. Additional Editor Comments: Thanks for submitting your revised paper to PLOS ONE, and I am pleased to inform you that your paper has now been approved by the previous peer expert. But before I can recommend the final editorial decision to our journal office, some minor issues need your attention. 1) Revise your title and specify the detailed study type or brief summary on the study contents. For example: Appropriate data segmentation improves speech encoding models: A comprehensive analysis of ....., or anything you think suitable. 2) Please consider to use structured abstract including background, objective, methods, results, and conclusion, and each of them shoule be presented as a separate paragraph, instead of just putting them together. 3) I note that there is no any table in the paper. Please consider to add some tables to summarize your core data, and it is better for readers to follow your main findings. [Note: HTML markup is below. Please do not edit.] 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 ******** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ****** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ****** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data 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 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 ****** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE 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 ****** 6. 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: Thank you for the opportunity to review the newly revised manuscript. I appreciate the description of the read speech, and my comments have all been successfully addressed. I have no further questions and wish the authors the best of luck with this project. ****** 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. 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 ******** [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.] 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. Registration is free. 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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0323276.r006
Revision 3
25 Mar 2025 Author Response Dear Dr. Yang, Thank you for your comments. We have added a summary to the title, it now reads: Appropriate data segmentation improves speech encoding models: Analysis and simulation of electrophysiological recordings. We also replaced ”Introduction” with ”Background” in the structured abstract and put each point on a separate paragraph. In your last point, you requested that we add a table to our publication to make it easier for the reader to follow the main findings. We believe that adding a table that compares the accuracy of the different encoding models with the different segment durations would likely be confusing and detracting rather than beneficial to readers. That is because the key point is the qualitative change in model accuracy that results from a change in segment duration, not the absolute values. We believe that the figures, combined with the text, illustrate this point well. Sincerely, Ole Bialas and Edmund Lalor Attachments Attachment Submitted filename: rebuttal3.pdf https://doi.org/10.1371/journal.pone.0323276.r007
26 Mar 2025 Decision Letter - Li Yang, Editor PONE-D-24-38627R3Appropriate Data Segmentation Improves Speech Encoding Models: Analysis and Simulation of Electrophysiological RecordingsPLOS ONE Dear Dr. Bialas, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’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 by May 10 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 plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ 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 academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. 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, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Li Yang, M.D. Academic Editor PLOS ONE Journal Requirements: Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice. Additional Editor Comments: Thanks for your response to my concerns. However, the abstract still can not meet the standard for publication. First, each section should be presented as a separate paragraph instead of just putting them together. Second, you should simplify the abstract since there are too many words included, such as background, methods, conclusion. Please just present the core points. Third, you should annotate the abbreviations in the abtract, such as Your current abstract in the manuscript: Background: In recent decades, studies modeling the neural processing of continuous, naturalistic, speech provided new insights into how speech and language are represented in the brain. However, the linear encoder models commonly used in such studies assume that the underlying data are stationary, varying to a fixed degree around a constant mean. Long, continuous, neural recordings may violate this assumption leading to impaired model performance. Objective: Our objective was to examine the effect of non-stationary trends in continuous neural recordings on the performance of linear speech encoding models. Methods: We used temporal response functions (TRFs), to predict continuous neural responses to speech while splitting the data into segments of varying length, prior to model fitting. Our Hypothesis was that, if the data were non-stationary, segmentation should improve model performance by making each segment approximately stationary. We used generative models for simulate and predict stationary and non-stationary neural data, testing our hypothesis under a known ground truth. We then used encoding models to predict the EEG recordings of participants who listened to a narrated story, testing our hypothesis on actual neural data. Results: Simulations showed that, for stationary data, increasing segmentation steadily decreased model performance. For non-stationary data however, segmentation initially improved model performance. Modeling of EEG recordings yielded similar results: segments of intermediate length (5-15 s) led to improved model performance compared to very short (1-2 s) and very long (30-120 s) segments. Conclusions: We showed that data segmentation improves the performance of encoding models in predicting both simulated and real neural data and that this can be explained by the fact that shorter segments approximate stationarity more closely. We thus conclude that the common practice of applying encoding models to long continuous segments of data is suboptimal and implore researchers to segment their data prior to model fitting. My recommend abstract format: Background: In recent decades, studies modeling the neural processing of continuous, naturalistic, speech provided new insights into how speech and language are represented in the brain. However, the linear encoder models commonly used in such studies assume that the underlying data are stationary, varying to a fixed degree around a constant mean. Long, continuous, neural recordings may violate this assumption leading to impaired model performance. Objective: We aimed to examine the effect of non-stationary trends in continuous neural recordings on the performance of linear speech encoding models. (This part can be merged into the introduction as well. Remember to delete the 'Objective:' if you choose to merge.) Methods: We used temporal response functions (TRFs), to predict continuous neural responses to speech while splitting the data into segments of varying length, prior to model fitting. Our Hypothesis was that, if the data were non-stationary, segmentation should improve model performance by making each segment approximately stationary. We used generative models for simulate and predict stationary and non-stationary neural data, testing our hypothesis under a known ground truth. We then used encoding models to predict the EEG recordings of participants who listened to a narrated story, testing our hypothesis on actual neural data. Results: Simulations showed that, for stationary data, increasing segmentation steadily decreased model performance. For non-stationary data however, segmentation initially improved model performance. Modeling of EEG recordings yielded similar results: segments of intermediate length (5-15 s) led to improved model performance compared to very short (1-2 s) and very long (30-120 s) segments. Conclusions: Our analysis showed that data segmentation improves the performance of encoding models in predicting both simulated and real neural data and that this can be explained by the fact that shorter segments approximate stationarity more closely. Thus, the common practice of applying encoding models to long continuous segments of data is suboptimal and implore researchers to segment their data prior to model fitting. [Note: HTML markup is below. Please do not edit.] [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.] 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. Registration is free. 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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0323276.r008
Revision 4
1 Apr 2025 Author Response Dear Dr. Yang, Thank you for providing an example of the desired formatting. We had inserted new paragraphs in the revised abstract, but the PLOS Latex template does not include a blank line after a paragraph. We changed the formatting of the abstract by manually inserting a blank line after every point. We also followed your suggestion to merge the ”Objectives" into the ”Background” section. We also made some minor changes for conciseness and hope that this meets the standards for publication. Sincerely, Ole Bialas and Edmund Lalor Attachments Attachment Submitted filename: rebuttal4.pdf https://doi.org/10.1371/journal.pone.0323276.r009
6 Apr 2025 Decision Letter - Li Yang, Editor Appropriate Data Segmentation Improves Speech Encoding Models: Analysis and Simulation of Electrophysiological Recordings PONE-D-24-38627R4 Dear Dr. Bialas, 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 will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. If you have any questions relating to publication charges, please contact our Author Billing department directly at authorbilling@plos.org. 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 onepress@plos.org. Kind regards, Li Yang, M.D. Academic Editor PLOS ONE Additional Editor Comments (optional): Thanks for the authors' efforts to comprehensively improve your manuscript according to editor's and reviewers' comments. I am pleased to inform you that your paper can be accepted for publication now. Thanks for the chance to assess your interesting and important work. Additionally, many thanks for all the reviewers' precious inputs. Reviewers' comments: https://doi.org/10.1371/journal.pone.0323276.r010
Formally Accepted
Acceptance Letter - Li Yang, Editor PONE-D-24-38627R4 PLOS ONE Dear Dr. Bialas, I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now being handed over to our production team. At this stage, our production department will prepare your paper for publication. This includes ensuring the following: * All references, tables, and figures are properly cited * All relevant supporting information is included in the manuscript submission, * There are no issues that prevent the paper from being properly typeset You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps. Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. If we can help with anything else, please email us at customercare@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. Li Yang Academic Editor PLOS ONE https://doi.org/10.1371/journal.pone.0323276.r011

Open letter on the publication of peer review reports

PLOS recognizes the benefits of transparency in the peer review process. Therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. Reviewers remain anonymous, unless they choose to reveal their names.

We encourage other journals to join us in this initiative. We hope that our action inspires the community, including researchers, research funders, and research institutions, to recognize the benefits of published peer review reports for all parts of the research system.

Learn more at ASAPbio .