Association between Neuroticism and Premenstrual Affective/Psychological Symptomatology

Round 1

Reviewer 1 Report

The study submitted for the review deals with clinically important problem of premenstrual negative changes in mood and behaviour, which concern a significant part of reproductive aged women population all over the world. Recognition of clinical significance of these mood and behaviour changes is currently confirmed by including premenstrual syndrome (PMS) and premenstrual dysphoric disorder (PMDD) into The Diagnostic and Statistical Manual of Mental Disorders. Numerous published studies attempted to identified risk factors and associations between temperamental/personality trait and occurrence of premenstrual mood and behaviour changes. This study fits into this attempts by investigating the association between personality trait neuroticism and occurrence of the selected seven affective and psychological symptoms. As mentioned by the authors such studies were conducted before and demonstrated mixed results. Thus "solid evidence" for this association is still needed to "expands opportunities for development of future fine-tuned diagnostic and treatment approaches".

Unfortunately, current study, due to critical problems with method description, data analysis and presentation of the results  does not provide such solid evidence. Below I'm listening my main concerns associated with the study.

1) Study sample - the sample is small containing only 45 women. Although the issue is potentially mitigated by observing women for 3 consecutive cycles it is unclear if data about symptoms from all of these cycles was included. My understanding from the description of the methods is that only one cycle was included. Exclusion criteria are not very clear for example they include exclusion based on having irregular menstrual cycle. It is unclear how this irregularity was defined. Similarly, prescribed medication is another exclusion criterium however it is unclear if any or specific medication was the reason for exclusion. On the other hand, the authors allowed for inclusion to study sample women with diagnosed anxiety and depression disorders. Since depression and anxiety are associated with presence of symptoms and clinical recognition of PMS and PMDD I'm curious about the consequences of this inclusion for the results of the study. Finally, more than half of the women in the sample had low and very low income when compared to overall population (median value of the income in 2020 around 41 000$). Again, as recognised by the authors in the first paragraph of the introduction, neuroticism is associated with low socioeconomic status. I'm curious about the consequences for the results of the study. I suggest that these factors (income, anxiety and depression) should be controlled for in their statistical analysis.

2) Anxiety and depression - data from Trait Anxiety and Depression Inventories were collected. I'm not sure how this data were incorporated in the analysis. 

3) Neuroticism - the Authors used The Zuckerman-Kuhlman Personality Questionnaire.  This short form measures the Zuckerman's five personality factors: Impulsive Sensation Seeking (ImpSS), Neuroticism-Anxiety (N-Anx), Aggression-Hostility (Agg-Host), Activity (Act), and Sociability (Sy). Yet, the authors claims assessing only neuroticism. Is there a way to separate Neuroticism from Anxiety in this questionnaire? Of note, there is at least three different "anxieties" assessed in the study - anxiety as the part of personality questionnaire, anxiety as a trait and anxiety as the symptom. It is hard to follow which anxiety is analysed where. Are this three somehow associated? Is anxiety as a trait increases anxiety as a symptom and as a part of neuroticism?

4) Social Adaptation Self-evaluation Scale (SASS) - how was this data incorporated in the analysis? What was the purpose to use this questionnaire in the context of the study?

5) Diagnosis of PMS and PMDD - unclear from the method section how the diagnosis was performed. Was it based on symptoms from 2 preceding cycles? Or only for the current cycle? Is the diagnosis (healthy, PMS, PMDD) somehow included in the main analysis? The authors identified control healthy group of participants. Was this control group somehow used in the main analysis or was it only used for basic comparisons?

6) Data analysis - I don't fully understand authors' use of statistical term "effect size". Effect size is a quantitative measure of the magnitude of the experimental effect. The larger the effect size the stronger the relationship between two variables. At least in two instances the authors use this term to describe dealing with the data at the level of individual variation. This should be avoided. Furthermore, I don't fully follow how the authors constructed the main generalized additive models. Was there seven separate models one for each symptom or one model including all seven symptoms. I suspect the first is true however it is not clear from the statistical analysis description. More detailed description of the statistical method together with the appropriate citation would be very useful in understanding authors statistical approach. The citation used in the current version of the manuscript refers the reader to the book "Rational choice in an uncertain world: The psychology of judgment and decision making". Is this book all devoted to the description of the generalized additive models? If not the specific part (chapters, pages) should be provided. Was there any confounders included in the model/s? As I mentioned before since the study set up is not experimental but rather observational there are several factors that may affect the results of the main associations. These factors should be controlled for.

7) Study results - the study sample consisted of 3 different groups of 1) healthy women; 2) women with PMS and 3) women with PMDD however the basic statistics are provided only for the entire sample. Including statistics and comparisons for 3 groups (or at least two - healthy vs diagnosed) would allow ensuring homogeneity of the sample and the reason for treating it as one in the further statistical analysis. Although, giving the complicated  psychological, physiological and brain structure and function associated symptomatology of diagnosed PMS and PMDD I'm not fully convinced that the group should be treated as one. It would also allow for identification of possible confounders. The caption for Figure 1 is Graphical presentation of Generalized Additive Modeling (GAM) for the Neuroticism - Anxiety Personality Table 0. and 0.0065, respectively. This information is very vague. What is Table 0. and 0.0065? What are the numbers on X axis on each panel. Are anxiety and depression presented in the figure premenstrual symptoms or psychological traits assessed with anxiety and depression questionnaires? What are the blue lines and grey areas presented on the figure? 

8) Theoretical issues - the authors claim that "Given the preliminary nature of the present study, we selected affective and psychological symptoms because of their hypothesized closer brain structural and functional proximity to personality traits compared to, for example, physical symptoms of PMS/PMDD" (page 4 lines 150 - 152)  What are the theoretical bases for this hypothesis? For example why the symptom of depression is included in the analysis and the symptom of lethargy, easy fatiguability and lack of energy, a main characteristics of depression not included in the analysis? The authors also claims that "Our study points that the relationship between anxiety and neuroticism does not appear to be mediated by shifting sex hormones (or their metabolites), while the relationship between low drive/interest and neuroticism does." (page 9 lines 273 -274). This is a very bold statement taking into consideration on the one hand the sample size of their study and on the other hand rich and solid literature suggesting that temperamental/personality traits and anxiety had a strong hormonal component.

The authors did not include any information about the ethical issues of the study e.g. no information about IRB or other ethical committee opinion about the study, no information about participants signing informed consent etc.

Author Response

REVIEWER 1 Response

1) Study sample - the sample is small containing only 45 women. Although the issue is potentially mitigated by observing women for 3 consecutive cycles it is unclear if data about symptoms from all of these cycles was included. My understanding from the description of the methods is that only one cycle was included. Exclusion criteria are not very clear for example they include exclusion based on having irregular menstrual cycle. It is unclear how this irregularity was defined. Similarly, prescribed medication is another exclusion criterium however it is unclear if any or specific medication was the reason for exclusion. On the other hand, the authors allowed for inclusion to study sample women with diagnosed anxiety and depression disorders. Since depression and anxiety are associated with presence of symptoms and clinical recognition of PMS and PMDD I'm curious about the consequences of this inclusion for the results of the study. Finally, more than half of the women in the sample had low and very low income when compared to overall population (median value of the income in 2020 around 41 000$). Again, as recognised by the authors in the first paragraph of the introduction, neuroticism is associated with low socioeconomic status. I'm curious about the consequences for the results of the study. I suggest that these factors (income, anxiety and depression) should be controlled for in their statistical analysis.

Response: We thank the reviewer for his/her comments. Data from all menstrual cycles was included in the analysis. To clarify this point, we added the following to the Data Analysis section (2.4): “For each DRSP symptom, we calculated the degree to which the symptom demonstrated an elevation in days -6 to -1 (“pre-menstruum”) from the start of the cycle relative to days +5 to +10 (“post-menstruum”). This was done for each woman for all available cycles.” Participants self-reported irregular menstrual cycles. This point is added in the exclusion criteria for clarity. Any prescription medication was exclusionary criteria. Although lifetime (but not current) anxiety and depression was permitted, none of the women included in this analysis reported lifetime incidence of anxiety or depression. The reason study participants reported low income is due to the fact that almost half of the participant sample consisted of students. We repeated the analysis including additional covariates. This is specified in response to question # 6.

2) Anxiety and depression - data from Trait Anxiety and Depression Inventories were collected. I'm not sure how this data were incorporated in the analysis. 

Response: As specified in the manuscript (Results section, a. Study Participants): “Our methodology of screening out participants with depression or anxiety seemed appropriate according to the results of STAI Y-2, BDI and SASS analyses, which showed normal score means and ranges.”

3) Neuroticism - the Authors used The Zuckerman-Kuhlman Personality Questionnaire.  This short form measures the Zuckerman's five personality factors: Impulsive Sensation Seeking (ImpSS), Neuroticism-Anxiety (N-Anx), Aggression-Hostility (Agg-Host), Activity (Act), and Sociability (Sy). Yet, the authors claims assessing only neuroticism. Is there a way to separate Neuroticism from Anxiety in this questionnaire? Of note, there is at least three different "anxieties" assessed in the study - anxiety as the part of personality questionnaire, anxiety as a trait and anxiety as the symptom. It is hard to follow which anxiety is analysed where. Are this three somehow associated? Is anxiety as a trait increases anxiety as a symptom and as a part of neuroticism?

Response: Neuroticism-anxiety is a single factor of The Zuckerman-Kuhlman Personality Questionnaire and cannot be separated. This is how neuroticism is operationalized in ZKPQ; however, generally, the literature does not refer to the trait as neuroticism-anxiety. For example, influential review papers (http://dx.doi.org/10.1016/j.neubiorev.2012.09.004 and http://dx.doi.org/10.1037/per0000274) refer to the trait as “neuroticism”. Neuroticism-anxiety as a trait (from ZKPQ) was assessed to identify which affective (including anxiety) and psychological premenstrual symptom(s) it associates with.

4) Social Adaptation Self-evaluation Scale (SASS) - how was this data incorporated in the analysis? What was the purpose to use this questionnaire in the context of the study?

Response: Altered social adaptation is a secondary effect that results from depressive symptoms (Otta et al., 2020; Journal of Affective Disorders). Social Adaptation Self-evaluation Scale (SASS) information was collected to assess accuracy of our procedure to screen out depression (by indirectly characterizing depressive symptomatology). The maximum score is 60, and the mean score for our sample was 46. As the higher scores indicate higher social adaptation; the high SASS score was suggestive of low depressive symptomatology. This was confirmed with a mean BDI score of 4.27, where scores 0-9 indicate minimal depression.

5) Diagnosis of PMS and PMDD - unclear from the method section how the diagnosis was performed. Was it based on symptoms from 2 preceding cycles? Or only for the current cycle? Is the diagnosis (healthy, PMS, PMDD) somehow included in the main analysis? The authors identified control healthy group of participants. Was this control group somehow used in the main analysis or was it only used for basic comparisons?

Response: The following statement is specified in the Data Analysis section: “The effect size greater than or equal to 1 reflects presence of a symptom [48] and we applied DSM-5 diagnostic criteria, as described in the “Introduction”.” As indicated in DSM-5, two or more prospective ratings of symptoms are necessary for establishing diagnosis, and this was applied in the current paper. Diagnosis was determined for descriptive purposes. As specified in the Data Analysis section, the analysis was completed for each individual symptom as a continuous variable. It is already known that PMDD patients score higher than healthy controls on neuroticism-anxiety; hence, this would not have been a novel finding. Rather, the objective of the analysis was to identify which affective and psychological symptoms associate with neuroticism-anxiety.

6) Data analysis - I don't fully understand authors' use of statistical term "effect size". Effect size is a quantitative measure of the magnitude of the experimental effect. The larger the effect size the stronger the relationship between two variables. At least in two instances the authors use this term to describe dealing with the data at the level of individual variation. This should be avoided. Furthermore, I don't fully follow how the authors constructed the main generalized additive models. Was there seven separate models one for each symptom or one model including all seven symptoms. I suspect the first is true however it is not clear from the statistical analysis description. More detailed description of the statistical method together with the appropriate citation would be very useful in understanding authors statistical approach. The citation used in the current version of the manuscript refers the reader to the book "Rational choice in an uncertain world: The psychology of judgment and decision making". Is this book all devoted to the description of the generalized additive models? If not the specific part (chapters, pages) should be provided. Was there any confounders included in the model/s? As I mentioned before since the study set up is not experimental but rather observational there are several factors that may affect the results of the main associations. These factors should be controlled for.

Response: The term “effect size” is taken from the literature describing the method premenstrual increase in symptomatology is calculated (i.e., “We subtracted the average post-menstruum score from the average pre-menstruum score and divided this score by participant-specific variance for each symptom. This essentially yielded an effect size for each woman and for each symptom”). This specific terminology is specified in Hartlage et al (2012) Archives of general psychiatry (please see reference section). We would like to keep the terminology as is in order to keep the literature to date consistent and prevent future confusion on this topic and hope that this decision is acceptable to the reviewer. We have updated the reference regarding GAM modeling with an article (Hastie and Tibshirani). In the revised version, we completed analysis comparing diagnostic groups on demographic characteristics and did not find any differences (please see new Supplementary Table 2). Moreover, we present results of the unadjusted model, as well as results of models adjusted for (1) age, and (2) age + age of menarche. Please see the revised version of Table 2, which presents results of adjusted models.

7) Study results - the study sample consisted of 3 different groups of 1) healthy women; 2) women with PMS and 3) women with PMDD however the basic statistics are provided only for the entire sample. Including statistics and comparisons for 3 groups (or at least two - healthy vs diagnosed) would allow ensuring homogeneity of the sample and the reason for treating it as one in the further statistical analysis. Although, giving the complicated  psychological, physiological and brain structure and function associated symptomatology of diagnosed PMS and PMDD I'm not fully convinced that the group should be treated as one. It would also allow for identification of possible confounders. The caption for Figure 1 is Graphical presentation of Generalized Additive Modeling (GAM) for the Neuroticism - Anxiety Personality Table 0. and 0.0065, respectively. This information is very vague. What is Table 0. and 0.0065? What are the numbers on X axis on each panel. Are anxiety and depression presented in the figure premenstrual symptoms or psychological traits assessed with anxiety and depression questionnaires? What are the blue lines and grey areas presented on the figure? 

Response: The following is the rationale for treating symptoms on a continuous scale, rather than classifying groups according to diagnostic criteria. To be diagnosed with PMDD, a woman needs to reach a symptom score of 1.0 on only one of the possible 4 affective symptoms, and she may or may not reach a score of 1.0 on a psychological symptom. For PMS, there needs to be only 1 symptom, which may or may not be affective and psychological in nature. This is described in detail in the “Introduction” section. Treating each symptom according to diagnostic group assignment may require a sample size of hundreds of participants as each cell would be diluted according to whether a woman has or has not met a particular symptom. Please see Supplementary Table 2 for comparison across demographic groups. The caption for Figure 1 reads:” Graphical presentation of Generalized Additive Modeling (GAM) for the Neuroticism-Anxiety Personality Trait in Relation to Premenstrual Symptoms. Psychological symptoms “low interest” and “difficulty concentrating” were significantly associated with neuroticism, with p values 0.0061 and 0.0065, respectively.”  The blue and gray lines represent fitted line and 95% interval, respectively.

8) Theoretical issues - the authors claim that "Given the preliminary nature of the present study, we selected affective and psychological symptoms because of their hypothesized closer brain structural and functional proximity to personality traits compared to, for example, physical symptoms of PMS/PMDD" (page 4 lines 150 - 152). What are the theoretical bases for this hypothesis? For example why the symptom of depression is included in the analysis and the symptom of lethargy, easy fatiguability and lack of energy, a main characteristics of depression not included in the analysis? The authors also claims that "Our study points that the relationship between anxiety and neuroticism does not appear to be mediated by shifting sex hormones (or their metabolites), while the relationship between low drive/interest and neuroticism does." (page 9 lines 273 -274). This is a very bold statement taking into consideration on the one hand the sample size of their study and on the other hand rich and solid literature suggesting that temperamental/personality traits and anxiety had a strong hormonal component.

Response: The theoretical basis for this hypothesis is that personality traits, like affective and psychological symptoms, are thought to originate from central sources. Premenstrual lethargy is considered a physical trait (please see reference Schmalenberger et al., 2017), which is considered lesser in proximity to CNS relative to affective and psychological symptoms.

The authors did not include any information about the ethical issues of the study e.g. no information about IRB or other ethical committee opinion about the study, no information about participants signing informed consent etc.

Response: The “Study Sample” section contains the following in the revised version: “Before any collection of data, study participants signed a consent form, approved by the University of Illinois Human Research Protection Office.”

Reviewer 2 Report

The present study sought to investigate the relationship between neuroticism and both affective and psychological premenstrual symptoms assessed with the daily record of severity of problems (DRSP) in 45 participants. Generalized additive modeling was used to describe relationships between neuroticism and the premenstrual increase in PMS/PMDD symptomatology. Significant associations were found for two of the psychological symptoms: low interest and difficulty concentrating. The authors conclude that the study improves our understanding of the premenstrual conditions and provides a platform for individualized treatment developments.

The introduction is well written, informative, and straightforward. Although globally clear, the methods section would benefit some important additional details, regarding the diagnosing procedure and the choice of DRSP items for example. Likewise, I suggest a few clarifications in the results section. The discussion could be toned down, considering the small sample size of the study. Interpretations and conclusions could be rephrased in a more temperate way. Some of the interpretations are problematic and should be elaborated on.

Main comments:

• Introduction, first paragraph: the first statement does not seem to be supported by the cited research (which is about assessing Intelligence, not including the definition of neuroticism). I suggest referring, for example, to the extensive review of Johan Ormel et al., 2013 about neuroticism (http://dx.doi.org/10.1016/j.neubiorev.2012.09.004), or equivalent work. Some helpful content can also be found in a more recent review from Howard S. Friedman, 2019 (http://dx.doi.org/10.1037/per0000274).

Accordingly, the first statement about neuroticism should specify that it does not concerns “emotions of all types”, as it is now stated, but specifically refers to negative emotions.

• Introduction, second paragraph: the reference #8 to early work is about the ovulatory phase, while the statement is about the premenstrual phase. The authors should modify the reference for the correct one: Benedek and Rubenstein. The correlations between ovarian activity and psychodynamic processes: II. The menstrual phase. Psychosomatic Medicine 1939, 1(4), 449-564.

• Methods, Study design: How were the blood and saliva samples used? Were hormones concentrations measured? If yes, additional analyses exploring the influence of hormone levels on the relationship between neuroticism and premenstrual symptoms could be informative.

• Methods, Study measures: As presented in Figure 1, the correspondence between the DSM-5 symptoms and the DRSP items always includes one single DRSP item per DSM symptom. However, the DRSP often includes more than one item for each DSM-5 symptom. For instance, the DSM affective lability symptom includes both the DRSP items “mood swings” and “sensitive to rejection/easily hurt”. Likewise, the DSM irritability/anger symptom includes “angry/irritable” and “conflicts”. The authors should provide the rationale for their choice of items (DRSP items 1, 5, 7 and 16).

In line with this, why doesn’t Supplementary Table 1 present all 21 DRSP items? Items 2, 3, 6, 8 and 17 are missing. There is also a typo indicating the “feeling depressed” item as the #8 item in the supplementary table 1, instead of #1 as in Table 1 and the Endicott (2006) reference.

• Methods, Data analysis: The authors state “The effect size greater than or equal to 1 reflects presence of a symptom [48] and we applied DSM-5 diagnostic criteria, as described in the “Introduction” (lines 189-190). One aspect of the diagnosis criteria for PMDD is missing here, as the symptoms in the premenstrual phase are required to be “marked”. Even though there is no consensus on which rating tool and thresholds to use for the diagnosis of PMDD, one must ensure, in addition to measuring the change in severity between the premenstrual and post menstruation phases, that the late luteal symptoms are at least moderate (for example > 4 using the DRSP scale), and that the mid-follicular symptoms are absent or minimal (for example < 2 using the DRSP). I realize that this might reduce even more the number of women meeting the criteria for PMDD in the present sample, but it is of upmost importance that the study participants are assigned to the correct diagnostic groups.

• Results, Study participants: Very little information is given about the three diagnostic groups, in terms of potential differences imparting bias in the subsequent analysis. Has any analysis been carried out to detect potential confounding variables? For example, any difference in age that could influence the differences in neuroticism?

• Results, Study participants: The authors write: “The analysis of neuroticism revealed that the PMDD group had higher neuroticism scores relative to the PMS and healthy control (p-adj=0.028 for both).” (lines 217-219). Can you specify to which one of the tests described in the method section the p-value relates? And what kind of adjustment was applied?

• Results, Table 2: I suggest the authors go through the “Updated Guidance on the Reporting of Race and Ethnicity in Medical and Science Journals”: https://jamanetwork.com/journals/jama/fullarticle/2783090. It is not clear whether the “race” information is relevant to neuroticism or premenstrual symptoms, and why/how it was collected.

• Discussions: It is not clear what is implied when the authors write: “The remaining affective symptomatology was also not associated with the trait despite an affective symptom requirement for PMDD diagnosis by DSM-5.” (lines 248-249). Can you elaborate?

There seem to be some confusion between the higher neuroticism scores found in women with PMDD compared to healthy controls, and the symptoms that define PMDD. While women with higher neuroticism might be more vulnerable to PMS and PMDD, this does not imply that neuroticism should be considered in the diagnosis criteria. Likewise, the fact that no association was found between neuroticism and affective symptoms in this sample does not undermine these as being core PMDD symptoms, especially considering the small number of participants in the PMDD group.

• Discussion: lines 273-275, can you clarify the following interpretation: “Our study points that the relationship between anxiety and neuroticism does not appear to be mediated by shifting sex hormones (or their metabolites), while the relationship between low drive/interest and neuroticism does.”?

In the present study, no association was found between premenstrual worsening of anxiety ratings and neuroticism scores. Hence, which relationship does the first part of the sentence refers to? It is not clear to which population this interpretation is applied (MDD, PMDD/PMS).

One may argue that the small number of women with PMDD included in the study could explain why no relationship with anxiety was found, as opposed to previous studies on MDD.  

Minor comments:

• Typo in the Figure 1 title/caption: “Figure 1. Graphical presentation of Generalized Additive Modeling (GAM) for the Neuroticism-Anxiety Personality Table 0. and 0.0065, respectively.” (line 241).

• Discussion: line 268, the abbreviation “MDD” is not specified. Although it is clear for experts in psychiatry that it refers to major depressive disorder, readers from another field might not make the connection with depression.

Author Response

REVIEWER 2 Response

The present study sought to investigate the relationship between neuroticism and both affective and psychological premenstrual symptoms assessed with the daily record of severity of problems (DRSP) in 45 participants. Generalized additive modeling was used to describe relationships between neuroticism and the premenstrual increase in PMS/PMDD symptomatology. Significant associations were found for two of the psychological symptoms: low interest and difficulty concentrating. The authors conclude that the study improves our understanding of the premenstrual conditions and provides a platform for individualized treatment developments.

The introduction is well written, informative, and straightforward. Although globally clear, the methods section would benefit some important additional details, regarding the diagnosing procedure and the choice of DRSP items for example. Likewise, I suggest a few clarifications in the results section. The discussion could be toned down, considering the small sample size of the study. Interpretations and conclusions could be rephrased in a more temperate way. Some of the interpretations are problematic and should be elaborated on.

Main comments:

Introduction, first paragraph: the first statement does not seem to be supported by the cited research (which is about assessing Intelligence, not including the definition of neuroticism). I suggest referring, for example, to the extensive review of Johan Ormel et al., 2013 about neuroticism (http://dx.doi.org/10.1016/j.neubiorev.2012.09.004), or equivalent work. Some helpful content can also be found in a more recent review from Howard S. Friedman, 2019 (http://dx.doi.org/10.1037/per0000274). Accordingly, the first statement about neuroticism should specify that it does not concerns “emotions of all types”, as it is now stated, but specifically refers to negative emotions.

Response: We thank the reviewer for providing his/her comments. The first section of the Introduction now reads: “Given its strong negative emotion basis, neuroticism is considered by some as the single most important factor associated with many forms of psychopathology and behavioral health [1] with important consequences on health and longevity [2].” The section incorporates both references provided by the reviewer.

Introduction, second paragraph: the reference #8 to early work is about the ovulatory phase, while the statement is about the premenstrual phase. The authors should modify the reference for the correct one: Benedek and Rubenstein. The correlations between ovarian activity and psychodynamic processes: II. The menstrual phase. Psychosomatic Medicine 1939, 1(4), 449-564.

Response: The reference has been updated in the revised version.

Methods, Study design: How were the blood and saliva samples used? Were hormones concentrations measured? If yes, additional analyses exploring the influence of hormone levels on the relationship between neuroticism and premenstrual symptoms could be informative.

Response: We agree that these analyses represent important considerations. Only a subset of women provided biological samples and, due to sample size limitations, we are presently unable to present results of these analyses. This is an ongoing study and hope to be able to do so in the future.

Methods, Study measures: As presented in Figure 1, the correspondence between the DSM-5 symptoms and the DRSP items always includes one single DRSP item per DSM symptom. However, the DRSP often includes more than one item for each DSM-5 symptom. For instance, the DSM affective lability symptom includes both the DRSP items “mood swings” and “sensitive to rejection/easily hurt”. Likewise, the DSM irritability/anger symptom includes “angry/irritable” and “conflicts”. The authors should provide the rationale for their choice of items (DRSP items 1, 5, 7 and 16).

Response: In order to be consistent in our approach, while limiting the number of comparisons, we selected the very first item on the list per affective DMS-5 symptom. For example, for marked affective lability (mood swings, feeling suddenly sad or tearful, or increased sensitivity to rejection), we selected mood swings. We clarify this on the bottom of Supplementary Table 1. Furthermore, the limitation section of the Discussion now reads: “We were not able to evaluate all DRSP items corresponding to a particular affective symptom (Supplementary Table 1). For example, affective lability could be operationalized as mood swings, feeling suddenly sad or tearful, or increased sensitivity to rejection, which represent different DRSP items. To limit the number of comparisons, while being consistent in our approach, we always selected the first item on the list of affective symptoms. For example, for affective lability, we analyzed only mood swings, but not increased sensitivity to rejection.”

In line with this, why doesn’t Supplementary Table 1 present all 21 DRSP items? Items 2, 3, 6, 8 and 17 are missing. There is also a typo indicating the “feeling depressed” item as the #8 item in the supplementary table 1, instead of #1 as in Table 1 and the Endicott (2006) reference.

Response: This information is included in the revised version.

Methods, Data analysis: The authors state “The effect size greater than or equal to 1 reflects presence of a symptom [48] and we applied DSM-5 diagnostic criteria, as described in the “Introduction” (lines 189-190). One aspect of the diagnosis criteria for PMDD is missing here, as the symptoms in the premenstrual phase are required to be “marked”. Even though there is no consensus on which rating tool and thresholds to use for the diagnosis of PMDD, one must ensure, in addition to measuring the change in severity between the premenstrual and post menstruation phases, that the late luteal symptoms are at least moderate (for example > 4 using the DRSP scale), and that the mid-follicular symptoms are absent or minimal (for example < 2 using the DRSP). I realize that this might reduce even more the number of women meeting the criteria for PMDD in the present sample, but it is of upmost importance that the study participants are assigned to the correct diagnostic groups.

Response: Indeed, the term “marked” in DSM-5 is unfortunately relative without a defined cut-off. We used a 100 percent increase in symptomatology while adjusting for total variance across the cycle from a published study (Hartlage et al., 2012), which provided empirical support for the number of items required for impairment and the temporal nature of symptom expression, bringing clinical significance into focus. The term marked was not incorporated into the analytic focus, perhaps due to its vagueness and difficulty what it means to an individual patient. This point is discussed in the limitations section: “We applied the analytical method of calculating effect size (i.e., symptom increase) from Hartlage et al (2012), which specifies that a 100% increase in premenstrual symptomatology is clinically relevant. However, this method does not necessitate that study participants were marking their symptoms as very severe (for example, 5 or 6 on the DRSP Likert scale).”

Results, Study participants: Very little information is given about the three diagnostic groups, in terms of potential differences imparting bias in the subsequent analysis. Has any analysis been carried out to detect potential confounding variables? For example, any difference in age that could influence the differences in neuroticism?

Response: The revised version provides this data (Supplemental Table 2), which shows no significant demographic differences across diagnostic groups. The revised version presents results of unadjusted models as well as models adjusted for current age and age of menarche. All models are corrected for multiple comparisons using the False Discovery Rate (FDR) method.

Results, Study participants: The authors write: “The analysis of neuroticism revealed that the PMDD group had higher neuroticism scores relative to the PMS and healthy control (p-adj=0.028 for both).” (lines 217-219). Can you specify to which one of the tests described in the method section the p-value relates? And what kind of adjustment was applied?

Response: This is located in the second paragraph of “Data Analysis” section: “We performed Kruskal-Wallis rank sum test, with the total neuroticism-anxiety score as the outcome and group (PMDD vs. PMS vs. healthy) as the predictor. We applied Wilcoxon rank sum tests to make pairwise comparisons between group levels with Benjamini-Hochberg corrections for multiple testing.”

Results, Table 2: I suggest the authors go through the “Updated Guidance on the Reporting of Race and Ethnicity in Medical and Science Journals”: https://jamanetwork.com/journals/jama/fullarticle/2783090. It is not clear whether the “race” information is relevant to neuroticism or premenstrual symptoms, and why/how it was collected.

Response: The following is included in the revised version: “Race was self-reported by study participants. Reporting race and ethnicity in this study was mandated by the US National Institutes of Health (NIH), consistent with the Inclusion of Women, Minorities, and Children policy. Individuals participating in the poststudy survey were categorized as American Indian or Alaska Native, Asian, Black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander, or White based on the NIH Policy on Reporting Race and Ethnicity Data.”

Discussions: It is not clear what is implied when the authors write: “The remaining affective symptomatology was also not associated with the trait despite an affective symptom requirement for PMDD diagnosis by DSM-5.” (lines 248-249). Can you elaborate?

Response: The paragraph is revised for clarity and now reads:” Results of the present study show that two premenstrual psychological symptoms – low interest and difficulty concentrating – are positively associated with the neuroticism personality trait. Contrary to our hypothesis, irritability/anger and anxiety did not demonstrate a significant association. The remaining affective symptomatology was also not associated with the trait despite an affective symptom requirement in DSM-5 for PMDD diagnosis. Hence, the strong relationship between neuroticism and PMDD does not appear to be mediated by, according to DSM-5, necessary affective symptomatology. Using stringent, prospective methodology to measure symptomatology, and a validated method to assess personality, the present study highlights co-expression of neuroticism with premenstrual decrease in interest and difficulty concentrating, which are psychological symptoms of PMDD. “

There seem to be some confusion between the higher neuroticism scores found in women with PMDD compared to healthy controls, and the symptoms that define PMDD. While women with higher neuroticism might be more vulnerable to PMS and PMDD, this does not imply that neuroticism should be considered in the diagnosis criteria. Likewise, the fact that no association was found between neuroticism and affective symptoms in this sample does not undermine these as being core PMDD symptoms, especially considering the small number of participants in the PMDD group.

Response: We have added the following to the first paragraph of the Discussion section: “Our findings do not imply that neuroticism should be considered in the diagnosis criteria. Rather, the study identifies neuroticism-based symptom expression in PMS/PMDD, thereby reflecting shared biological mechanisms and expanding opportunities for development of future clinical approaches.”

Discussion: lines 273-275, can you clarify the following interpretation: “Our study points that the relationship between anxiety and neuroticism does not appear to be mediated by shifting sex hormones (or their metabolites), while the relationship between low drive/interest and neuroticism does.”?

Response: This statement highlights the difference in associations found between studies evaluating MDDsymptoms-neuroticism and PMDD/PMS symptoms-neuroticism. As PMS/PMDD are conditions of hormonal etiology, finding a relationship between PMDD/PMS symptoms-neuroticism points hormonal underpinnings, which is not necessarily the case for MDD symptom-neuroticism association identification.

In the present study, no association was found between premenstrual worsening of anxiety ratings and neuroticism scores. Hence, which relationship does the first part of the sentence refers to? It is not clear to which population this interpretation is applied (MDD, PMDD/PMS).

Response: This sentence is revised for clarity and now reads: “Results of our study, therefore, suggest that the relationship between low drive/interest and neuroticism appears to have a hormonal basis, while, the relationship between anxiety and neuroticism does not, as MDD is not considered a disorder caused by shifting sex hormones (or their metabolites).”

One may argue that the small number of women with PMDD included in the study could explain why no relationship with anxiety was found, as opposed to previous studies on MDD.  

Response: The Discussion section (paragraph where MDD and PMDD are compared) now points this as a limitation. The following sentence is included: “Our findings, however, will need to be replicated in a study involving a larger sample size.”

Minor comments:

Typo in the Figure 1 title/caption: “Figure 1. Graphical presentation of Generalized Additive Modeling (GAM) for the Neuroticism-Anxiety Personality Table 0. and 0.0065, respectively.” (line 241).

 Response: This is corrected.

Discussion: line 268, the abbreviation “MDD” is not specified. Although it is clear for experts in psychiatry that it refers to major depressive disorder, readers from another field might not make the connection with depression.

Response: This is corrected.

Round 2

Reviewer 1 Report

Thank you for discussing all raised issues. I don't have any additional comments. 

Author Response

Thanks for your comments

Reviewer 2 Report

The authors have addressed my comments in a satisfactory manner. 

Author Response

Thanks for your review work