Emotion dynamic patterns between intimate relationship partners predict their separation two years later: A machine learning approach

Peter Hilpert; Matthew R. Vowels; Merijn Mestdagh; Laura Sels

doi:10.1371/journal.pone.0288048

Peer Review History

Original SubmissionDecember 13, 2022
6 Mar 2023 Decision Letter - Joydeep Bhattacharya, Editor PONE-D-22-33994Emotion Dynamic Patterns Between Intimate Relationship Partners Predict Their Separation Two Years Later: A Machine Learning ApproachPLOS ONE Dear Dr. Hilpert, 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 Apr 20 2023 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, Joydeep Bhattacharya 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 provide additional details regarding ethical approval in the body of your manuscript. In the Methods section, please ensure that you have specified the name of the IRB/ethics committee that approved your study 3. Peer review at PLOS ONE is not double-blinded (https://journals.plos.org/plosone/s/editorial-and-peer-review-process). For this reason, authors should include in the revised manuscript all the information removed for blind review 3. Thank you for stating the following financial disclosure: "EC, PK: GOA/15/003; OT/11/031, Research Fund of the University of Leuven, https://www.kuleuven.be/english/research/support/if EC, PK: IAP/P7/06, Interuniversity Attraction Poles programme, http://www.belspo.be/belspo/iap/index_en.stm EC, PK, FT: G.0582.14, Fund for Scientific Research-Flanders, https://www.fwo.be/en/ PH: P2ZHP1_151628, Swiss National Science Foundation, https://www.snf.ch PH: P300P1_164582, Swiss National Science Foundation, https://www.snf.ch PH: P3P3P1_174466, Swiss National Science Foundation, https://www.snf.ch". Please state what role the funders took in the study. If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript." If this statement is not correct you must amend it as needed. Please include this amended Role of Funder statement in your cover letter; we will change the online submission form on your behalf. 4. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide. 5. 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: I had trouble finding suitable referees which explains the delay in getting to a decision. The one report I have in possession is very positive about the paper. My own quick reading also left me with a positive impression. I encourage you to revise the paper based off the comments of the referee which seem doable (esp. the part about showing circumstantial evidence about causation). [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: Yes ******** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ****** 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 ****** 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 ****** 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: The paper addresses a relevant question and is methodologically sound. I would recommend accepting the paper provided the following revisions are adequately addressed. Major The paper would benefit from showing additional benchmarks. For example, it is likely that demographic features and their interactions are predictive of relationship outcomes. Other information that was collected about participants could also be used to construct theory informed benchmarks. It would be interesting to see whether the emotion features would incrementally improve predictive performance when added to the same model (i.e. whether they capture non-redundant target variance). Being able to show that the emotion features capture non-redundant information would make the paper much stronger. The directionality of the association between emotion variability and relationship stability remains unclear and the authors suggest this as a direction for future research. I strongly recommend addressing this question in the current paper. This could easily be done by showing descriptive relationships between properties of the distribution of input features and the target variable. For example, the authors could compute point-biserial correlations between the person-level variance scores of emotion measures and the relationship outcome. For completeness this could also be done for the mean. Alternatively, the authors could show differences between the distributions of relationship outcome classes the same way it was done for CPSD (fig 4). Minor MCC is a relatively uncommon measure for classification performance. For completeness the authors should add the standard metrics (accuracy, precision, recall, and AUC) to the visualizations and consider reporting AUC as the main metric throughout the paper. The authors refer to their study as an experimental study on several occasions (e.g. page 21). This is imprecise as the study is clearly correlational and there was no randomly assigned manipulation. The authors should correct this. Formatting / style The following sentence is incomplete: “After classifying 10-second sequences as positive, negative, or neutral, they predicted partners’ next sequence the person’s prior sequence and found that unhappy couples experience greater reciprocity of negative affect (17,24) and high affect reciprocity 75 (12) than happy couples.” The title on page 1 is different from the title on the cover page ****** 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 ******** [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.0288048.r001
Revision 1
15 Jun 2023 Author Response PONE-D-22-33994 Dear editor and dear reviewer, Thank you very much for the suggestions. We re-run all the analysis and added the sociademographic variables, which were availeable to us (this was limited as this is a secondary data analysis). We tried to implement the comments as good as possible and hope that this has improved the manuscript. Review Comments to the Author Reviewer #1: The paper addresses a relevant question and is methodologically sound. I would recommend accepting the paper provided the following revisions are adequately addressed. COMMENT 1 Major The paper would benefit from showing additional benchmarks. For example, it is likely that demographic features and their interactions are predictive of relationship outcomes. Other information that was collected about participants could also be used to construct theory informed benchmarks. It would be interesting to see whether the emotion features would incrementally improve predictive performance when added to the same model (i.e. whether they capture non-redundant target variance). Being able to show that the emotion features capture non-redundant information would make the paper much stronger. ANSWER 1 Thank you for this very interesting question. In fact, creating a comparison standard is an excellent way to test whether the variability in emotions plays a role or whether emotions could potentially be a proxy for other factors. Therefore, we first calculated all models with the additional variables age (of each partner in the couple) and relationship living status (married, cohabiting, not cohabiting) to compare whether there was any difference, with the idea that if there were differences, we could proceed step by step to see which factors improve performance. More variables could not be included as this is a secondary data analysis and these are all the socioeconomic data we have access to. However, the results show that the additional variables do not lead to a change in the results. In fact, we see some reduction in predictive performance. Whilst in our experience it is more common for the addition of information to improve predictive performance, we have sometimes found that, for datasets with small effects, the relationship is not always that of a monotonic improvement. In any case, the results indicate that the effect is explained by emotional experience and dynamics between partners, and that emotions contain non-redundant information. It could have been the case that emotions were only a proxy for other factors, but the additional analyses indicate that this is not the case. Here are all the examples, which will be included into the supplementary material (on the right sight all are the figures including additionally the two demographic variables. Unfortunately, we could not include the plots here. But the plots and the comparison can be found in the supplementary material A) Positive interactions versus conflict interactions Random Forest MMC vs Random Forest MMC & demographics Random Forest AUC vs Random Forest AUC & demographics Overall, it appears that adding demographic variables has resulted in a slight decline in prediction accuracy, rather than an improvement. B) Figures for Positive interactions Logistic Regression vs Logistic Regression & demographics Random Forest vs Random Forest & demographics Random Forest & AUC vs Random Forest & AUC & demographics C) Figures for Conflict interactions Logistic Regression vs Logistic Regression & demographics Random Forest vs Random Forest & demographics Random Forest & AUC vs Random Forest & AUC & demographics It is clear that the demographic variables had no significant influence on the results. All these plots can be found in the supplemental materials. This reinforces our confidence in the findings. It is actually quite remarkable that emotional information from just 10 minutes of interaction can predict separation two years later. We would not argue that the observed emotions cause separation, but it may serve as a proxy for how partners interact and experience each other, which is indicative of a potential future separation, and this relationship seems worthy of future replication and exploration. We have added this clarification in the article. Statistical Analysis “As it is possible that emotional experience and dynamics between partners may capture redundant information, all models are compared with extended models that include demographic factors as predictors. This helps us determine whether emotions may capture redundant information.” Results “Finally, all models were then compared to models in which we included additional information regarding demographics. However, since there were no significant differences, these results have been included as supplemental material.” COMMENT 2 The directionality of the association between emotion variability and relationship stability remains unclear and the authors suggest this as a direction for future research. I strongly recommend addressing this question in the current paper. This could easily be done by showing descriptive relationships between properties of the distribution of input features and the target variable. For example, the authors could compute point-biserial correlations between the person-level variance scores of emotion measures and the relationship outcome. For completeness this could also be done for the mean. Alternatively, the authors could show differences between the distributions of relationship outcome classes the same way it was done for CPSD (fig 4). ANSWER 2 Thank you for this idea. We have already attempted to go beyond pure prediction and understand conceptually whether more or less variance predicts separation. This would be interesting to know because exploratively we could draw the conclusion whether too much emotional fluctuation or too little fluctuation might have something to do with later separation, even though this would only be on an exploratory level. As suggested, we have examined the biserial correlations to see if there are any relationships between properties of the distribution of input features and the target variable. However, the results of the biserial correlations indicate no significant relationship between the variance of someone's time series and the breakup. All p-values are greater than 0.05. The result supports our previous statement that the reason for the relationship is unclear, and that further research is needed. We have supplemented the manuscript with the table of correlations and provided a more detailed description. «This indicates that the variability of emotions across the interaction did contain important information about whether or not couples will break up. As the variance information is at the individual level, rather than the couple level, it is non-trivial to identify whether, for instance, higher or lower variance is associated with breakup. For example, it may be that partner A has high variance, and partner B has low variance. Additional point-biserial correlations were computed between the individual emotion variability over a sequence and the breakup (Table 1). As the correlations were not significant, the direction of the relationship is not clear, indicating that variability alone is not sufficient in determining breakup. Further research is needed to specifically investigate the nature of the relationship between emotion variance and breakup. Table 1. Point-biserial correlations examining whether the emotional variance of someone's time series correlates with whether they break up. Variance Break-up p-value Male, positive interaction 0.10 .310 Female, positive interaction -0.06 .567 Male, conflict interaction 0.18 .086 Female, conflict interaction 0.16 .122 MINOR COMMENTS MCC is a relatively uncommon measure for classification performance. For completeness the authors should add the standard metrics (accuracy, precision, recall, and AUC) to the visualizations and consider reporting AUC as the main metric throughout the paper. ANSWER: We agree with the reviewer that for individuals familiar with machine learning, AUC is the better approach (amongst others which the reviewer mentioned) to measure classification performance. However, our experience has shown that psychologists are often unfamiliar with AUC, as it is not part of the standard statistical training. Although MMC is also not a standard tool in the psychologist's repertoire, we have found that psychologists have a better understanding of MMC, as they are familiar with the concept of correlation. Therefore, to ensure the results are understandable for psychologists, we will maintain MMC in the manuscript. Nevertheless, we have also calculated the standard metric of AUC and present it in the supplemental materials. COMMENT The authors refer to their study as an experimental study on several occasions (e.g. page 21). This is imprecise as the study is clearly correlational and there was no randomly assigned manipulation. The authors should correct this. ANSWER: In our opinion, it is an experiment because we have a manipulation - the interactions (positive interaction versus conflict interaction). Yes, there is no control group of other couples in this sense, but we can now compare the two interactions. This means that the control group for the positive interaction is the data from the conflict interaction and vice versa. This is the reason why we also compared then in this study. We believe that this is a very good control group and better than comparing it with other couples. Therefore, it is not actually an observation since we manipulated the interactions. We can change the terminology if the reviewer insists, but in our opinion, the term "observational data" is not correct in this case since we "forced" people into different interactions. COMMENT Formatting / style ANSWER: We have reformatted the manuscript and made it compatible with PLOS ONE guidelines. COMMENT The following sentence is incomplete: “After classifying 10-second sequences as positive, negative, or neutral, they predicted partners’ next sequence the person’s prior sequence and found that unhappy couples experience greater reciprocity of negative affect (17,24) and high affect reciprocity 75 (12) than happy couples.” ANSWER: We revised the sentence: ” After classifying 10-second sequences as positive, negative, or neutral, the researchers predicted their partners' next sequence based on their prior sequence. They found that unhappy couples exhibit greater reciprocity of negative affect [17], [24] and higher affect reciprocity compared to happy couples [12].” COMMENT The title on page 1 is different from the title on the cover page ANSWER: Title: Emotion Dynamic Patterns Between Intimate Relationship Partners Predict Their Separation Two Years Later: A Machine Learning Approach Attachments Attachment Submitted filename: Letter_Reviewer.docx https://doi.org/10.1371/journal.pone.0288048.r002
19 Jun 2023 Decision Letter - Joydeep Bhattacharya, Editor Emotion Dynamic Patterns Between Intimate Relationship Partners Predict Their Separation Two Years Later: A Machine Learning Approach PONE-D-22-33994R1 Dear Dr. Hilpert, 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. Congratulations on a fine paper! 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 http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, 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, Joydeep Bhattacharya Academic Editor PLOS ONE Additional Editor Comments (optional): Reviewers' comments: https://doi.org/10.1371/journal.pone.0288048.r003
Formally Accepted
23 Jun 2023 Acceptance Letter - Joydeep Bhattacharya, Editor PONE-D-22-33994R1 Emotion dynamic patterns between intimate relationship partners predict their separation two years later: A machine learning approach Dear Dr. Hilpert: I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department. 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 plosone@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. Joydeep Bhattacharya Academic Editor PLOS ONE https://doi.org/10.1371/journal.pone.0288048.r004

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 .