Disruption of Circadian Sleep/Wake Rhythms in Infants May Herald Future Development of Autism Spectrum Disorder

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you very much for the opportunity to review this paper. This study presents an interesting and important topic. However, some questions need to be resolved in order to consider this manuscript ready to publication:

 

1. Keywords: it is suggested to delete the numbers contained in the keywords. It is not necessary to enumerate them but only to cite.

2. Lines 42-43 says “wake-up time after 07:00h and/or 08:00h on weekdays and holidays”. Are the authors really referring to holidays or would the correct word be “weekends”?

3. Line 51 says “The prevalence of autism spectrum disorder (ASD) is considered to be increasing …”. In this regard, is the prevalence of ASD really increasing and/or would these cases be better identified today?

4. Lines 86-90 says “In the present study, we would like to express new concepts about ASD from a chronobiological viewpoint and would like to discuss a possible treatment strategy”. Are the concepts presented by the authors new or have they already been discussed in the literature? Please, confirm or clarify this information.

5. In the Discussion section, the authors have successfully presented the importance of this work. It is suggested that some of this information be briefly included in the Introduction section as well.

6. In Table 1, it is suggested to increase the font size and, if possible, to include the percentage next to the absolute number, between parenthesis.

7. There are two captions corresponding to figure 1 and two captions corresponding to figure 2 with repeated information. Please, correct them.

8. In the Introduction section, lines 59-64, the authors describe the “apathetic type”. Is it possible that “apathetic type” infants wake up at night and stay awake for > 1 hour without their parents noticing? Would some of the characteristics of sleep disturbance be biased / underestimated for this type?

9. In the Results section, line 136, it is suggested that the word “rats” be changed to “prevalence” or “proportion”.

10. The Abstract says “We investigated if the abnormal sleep-wake rhythms in infants are related to the future development of autism spectrum disorder (ASD)”. Based on that, I think the goal of this work was to compare two groups: ASD vs control. However, in the present analysis, the authors chose to create two multiple logistic models (control vs. K-ASD groups and control vs. H-ASD groups). Why the authors did not combine K-ASD and H-ASD into a single group and compare this combined group to the control group is not clear to me. If such a separation is necessary, would not a comparison of the K-ASD and the H-ASD be an interesting result? Please clarify.

11. To construct multiple logistic models, would it not be interesting for the authors to combine the classes with low prevalence? For example, for the bedtime variable, consider the classes "before 21h", "21h-22h", ">22h" and for the waking time variable, consider the classes "before 7h", "7h-8h" and ">8h". This alternative would provide more accurate confidence intervals for the OR.

12. In the Section 2.2, lines 113-114, it is sad that “children in the control group in infancy, at 1-3≦ (listed as 1-3 below) years of age and at 3-5≦ (listed as 3-5 below) years of age”. Which age group does the “infancy” group correspond to? Also, in Table 1, there are ages ranging from 2 to 6 years. Which group were the 5 years-old and 6 years-old children assigned to? Furthermore, according to table 1, there are 38 children belonging to control group and age 1-3≦ (= 2 y.o.) while, in table 2, where counted 142 children in that same group. Would these 2 counts be supposed to be equal? Please, clarify.

13. The caption of Tables 2 to 5 states that the ORs were adjusted for gender. Was gender included as an explanatory variable in the model? Could the authors, please, clarify why a Bonferroni correction was applied to a multiple logistic regression.

14. When proposing a multiple logistic analysis it is appropriate to assess the goodness of fit beyond the Hosmer-Lemeshow test,  using measures such as AUC, sensitivity, specificity, error rate and others. These measure whether the model is good at prediction.

15. Lines 165-167 says “Waking times of 7:00 to 8:00 and 8:00 to 9:00 on weekdays and on holidays, and after 9:00 on holidays in infancy and at 1-3 years of age were associated with the prevalence of ASD (Tables 4 and 5)”. However, this information in bold does not seem to me to be correct for the H-ASD group (Table 5). Could you please confirm this information?

Comments on the Quality of English Language

Extensive English language editing is required.

Author Response

Thank you very much for your kind comments and suggestions for reviewing this document. We have revised the current version according to your comments. Below are your suggestions and our responses. We hope that it has been corrected appropriately. 

 

1. Keywords: it is suggested to delete the numbers contained in the keywords. It is not necessary to enumerate them but only to cite.

Thank you for pointing out this issue. The number has been deleted.

 

2. Lines 42-43 says “wake-up time after 07:00h and/or 08:00h on weekdays and holidays”. Are the authors really referring to holidays or would the correct word be “weekends”?

 

Thank you for pointing out this issue. We corrected “holiday” to “weekend” in the revised manuscript.

 

3. Line 51 says “The prevalence of autism spectrum disorder (ASD) is considered to be increasing …”. In this regard, is the prevalence of ASD really increasing and/or would these cases be better identified today?

 

Based on previous reports and information from the semesters in Japan, the majority opinion is that the number of 　ASD　cases is clearly increasing.

 

 

4. Lines 86-90 says “In the present study, we would like to express new concepts about ASD from a chronobiological viewpoint and would like to discuss a possible treatment strategy”. Are the concepts presented by the authors new or have they already been discussed in the literature? Please, confirm or clarify this information.

 

Added the following text　Lines　95:　In this study, we would like to express a new concept regarding ASD from a chronobiological perspective that has been discussing in recent years [10,11,15,16], and discuss possible treatment strategies.

 

5. In the Discussion section, the authors have successfully presented the importance of this work. It is suggested that some of this information be briefly included in the Introduction section as well.

 

We added the following text to the Introduction according to comments.　Actually, it has been suggested that biological rhythm disorders, including the rhythm in ASD, are not only associated symptoms, but are also deeply related to its pathogenesis [15-18].  More importantly, the circadian rhythm controls the so-called life-support systems (e.g., the autonomic nervous, hormone secretion, thermoregulation, etc.) [13,16]. Therefore, the chrono-biological theory is logical and convincing for understanding and/or explaining the various clinical manifestations, which spread over the whole body of ASD [15,16,19].

 

6. In Table 1, it is suggested to increase the font size and, if possible, to include the percentage next to the absolute number, between parenthesis.

 

As suggested by the reviewer, we have increased the font size and added percentages in Table 1.

 

7. There are two captions corresponding to figure 1 and two captions corresponding to figure 2 with repeated information. Please, correct them.

 

Thank you for pointing out this issue. We corrected the captions in the revised manuscript.

 

8. In the Introduction section, lines 59-64, the authors describe the “apathetic type”. Is it possible that “apathetic type” infants wake up at night and stay awake for > 1 hour without their parents noticing? Would some of the characteristics of sleep disturbance be biased / underestimated for this type?

 

Although we cannot deny the possibility of children being awake without their parents noticing, in many cases they will notice, so we think it is unlikely that their children will be awake for more than an hour. It is known that some children of the apathetic type later become the normal type and some children become the irritable type, but we have not investigated the details of this　points.

 

9. In the Results section, line 136, it is suggested that the word “rats” be changed to “prevalence” or “proportion”.

Thank you for pointing out this issue. We corrected the captions in the revised manuscript.

 

10. The Abstract says “We investigated if the abnormal sleep-wake rhythms in infants are related to the future development of autism spectrum disorder (ASD)”. Based on that, I think the goal of this work was to compare two groups: ASD vs control. However, in the present analysis, the authors chose to create two multiple logistic models (control vs. K-ASD groups and control vs. H-ASD groups). Why the authors did not combine K-ASD and H-ASD into a single group and compare this combined group to the control group is not clear to me. If such a separation is necessary, would not a comparison of the K-ASD and the H-ASD be an interesting result? Please clarify.

Since the data were not from the same organization, we considered whether it would be appropriate to mix them together, but we expected to obtain almost similar results for the two organizations, so I considered them separately.

 

11. To construct multiple logistic models, would it not be interesting for the authors to combine the classes with low prevalence? For example, for the bedtime variable, consider the classes "before 21h", "21h-22h", ">22h" and for the waking time variable, consider the classes "before 7h", "7h-8h" and ">8h". This alternative would provide more accurate confidence intervals for the OR.

. Thank you for your valuable suggestion. According to the reviewer’s suggestion, we reclassified the bedtimes into the classes "before 21:00," "21:00-22:00," and " after 22:00" and the waking times into the classes "before 7:00," "7:00-8:00," and " after 8:00" for the analysis. As a result, we were able to provide more accurate 95% confidence intervals for the ORs. We revised the Abstract section, Results section, and Tables 2-5 in the manuscript.

 

 

12. In the Section 2.2, lines 113-114, it is sad that “children in the control group in infancy, at 1-3≦ (listed as 1-3 below) years of age and at 3-5≦ (listed as 3-5 below) years of age”. Which age group does the “infancy” group correspond to? Also, in Table 1, there are ages ranging from 2 to 6 years. Which group were the 5 years-old and 6 years-old children assigned to? Furthermore, according to table 1, there are 38 children belonging to control group and age 1-3≦ (= 2 y.o.) while, in table 2, where counted 142 children in that same group. Would these 2 counts be supposed to be equal? Please, clarify.

Table 1 shows the age distribution of the children at the time of the survey, and Table 2 shows the situation from infancy to age 5 from a retrospective perspective, so the numbers are different. The title of Table 1 has been slightly changed in the revised manuscript for clarity.

 

13. The caption of Tables 2 to 5 states that the ORs were adjusted for gender. Was gender included as an explanatory variable in the model? Could the authors, please, clarify why a Bonferroni correction was applied to a multiple logistic regression.

 

In the multivariable logistic regression models, the ORs were adjusted by adding sex as an explanatory variable. The Bonferroni correction for multivariable logistic regression analyses were removed. The above statement is reflected in the Materials and Methods section and Tables 2-5 of the revised manuscript.

 

14. When proposing a multiple logistic analysis it is appropriate to assess the goodness of fit beyond the Hosmer-Lemeshow test,  using measures such as AUC, sensitivity, specificity, error rate and others. These measure whether the model is good at prediction.

 

Following the reviewer's suggestion, we performed the Hosmer-Lemeshow test using the χ squared distribution. As a result, P-values were greater than 0.05 for all multivariate logistic regression models, confirming the predictive ability of the models. We added the above statement in the Materials and Methods section, Results section, and Tables 2-5 of the revised manuscript.

 

15. Lines 165-167 says “Waking times of 7:00 to 8:00 and 8:00 to 9:00 on weekdays and on holidays, and after 9:00 on holidays in infancy and at 1-3 years of age were associated with the prevalence of ASD (Tables 4 and 5)”. However, this information in bold does not seem to me to be correct for the H-ASD group (Table 5). Could you please confirm this information?

 

Thank you for pointing out this issue. We corrected this point in the results section of revised manuscript.

 

Reviewer 2 Report

Comments and Suggestions for Authors

This is an interesting study. It may also advance the current knowledge given the concerns associated with predicting ASD. However, the manuscript is not well-written in general, and deficiencies of some sections makes comprehending the other sections difficult. Please see below my more specific comments.

 

Introduction

The introduction section is poorly organized and not focussed. This section could be written more concisely and with a better focus, clearly articulating the aim of the study. Without a clear understanding of the aim, commenting on the other section of the manuscript is difficult.

 

Methods

The methodological details are very difficult to understand. It also lacks adequate details to understand the contexts, the nature of the institutions included and the nature of the participants in these institutions and study locations. This section must clearly identify the study design, the study area/institutions, the study population, the sample and the sampling process and the data collection/extraction/review procedures.

This insufficiency of methodological details makes making valid comments on the results and the discussion sections difficult.

 

Results

The authors need to be careful of the use of technical jargon. For example, ‘rate’ has a time dimension and has been used to describe the results, but from the methodological details that have been provided, it does not seem that there was a time element.

 

Discussion

I have not made any comments on the discussion section as I do not believe that I can make any valid comments given the difficulty in understanding the methods and results.

Comments on the Quality of English Language

The use of language is satisfactory, but some editing may be required. 

Author Response

 Thank you very much for your kind comments and suggestions for reviewing this document. We have revised the current version according to your comments.

Introduction:  The introduction section is poorly organized and not focussed. This section could be written more concisely and with a better focus, clearly articulating the aim of the study. Without a clear understanding of the aim, commenting on the other section of the manuscript is difficult.

We have added the following sentence to the introduction section to clarify the purpose of our study.　

 Actually, it has been suggested that biological rhythm disorders, including the rhythm in ASD, are not only associated symptoms, but are also deeply related to its pathogenesis [15-18].  More importantly, the circadian rhythm controls the so-called life-support systems (e.g., the autonomic nervous, hormone secretion, thermoregulation, etc.) [13,16]. Therefore, the chrono-biological theory is logical and convincing for understanding and/or explaining the various clinical manifestations, which spread over the whole body of ASD [15,16,19]. According to the principles of early diagnosis and early treatment, we investigated whether the abnormal rhythms in infants are related to the future development of autism spectrum disorder (ASD).

 

Materials and Methods：

The methodological details are very difficult to understand. It also lacks adequate details to understand the contexts, the nature of the institutions included and the nature of the participants in these institutions and study locations. This section must clearly identify the study design, the study area/institutions, the study population, the sample and the sampling process and the data collection/extraction/review procedures.

This insufficiency of methodological details makes making valid comments on the results and the discussion sections difficult.

The data were collected from different regions, and the reason we did not mix the data was because we thought it was important to obtain similar results in different regions. In the revised version, the methodology and analysis descriptions have been revised for clarity. The above statement is reflected in the Materials and Methods section and Tables 2-5 of the revised manuscript.

Results

The authors need to be careful of the use of technical jargon. For example, ‘rate’ has a time dimension and has been used to describe the results, but from the methodological details that have been provided, it does not seem that there was a time element.

We have revised the description in the results section following reviewer comments.

 

 

 

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I would like to thank the authors for answering my questions, accepting the main suggestions and making the needed changes in the manuscript. However, some questions still need to be resolved in order to consider this manuscript ready for publication:

 

1. Table 1 is placed in the middle of the paragraph. Please, move this table to the end of this paragraph.

2. In line 109, it is suggested to insert “According to Figure 1, in both the ASD groups,...”

3. It is suggested that the sentence in lines 131-134 “Having bedtimes of after 22:00 on weekdays and on weekends at infancy were associated with the prevalence of ASD (Tables 2 and 3). Having bedtimes of after 22:00 on weekdays and on weekends at 1-3 years of age were associated with the prevalence of ASD (Tables 2 and 3)” be rewritten as follows “Having bedtimes of after 22:00 on weekdays and on weekends at infancy and at 1-3 years of age were associated with the prevalence of ASD (Tables 2 and 3).”

 

Author Response

Thank you for your appropriate advice.

The manuscript has been revised according to the instructions of the commenters below.

1. Table 1 is placed in the middle of the paragraph. Please, move this table to the end of this paragraph.

Table 1 has been moved after the Results section (after line 114).

 

1. In line 109, it is suggested to insert “According to Figure 1, in both the ASD groups,...”

“According to Figure 1”  is added on line 104.

 

1. It is suggested that the sentence in lines 131-134 “Having bedtimes of after 22:00 on weekdays and on weekends at infancy were associated with the prevalence of ASD (Tables 2 and 3). Having bedtimes of after 22:00 on weekdays and on weekends at 1-3 years of age were associated with the prevalence of ASD (Tables 2 and 3)” be rewritten as follows “Having bedtimes of after 22:00 on weekdays and on weekends at infancy and at 1-3 years of age were associated with the prevalence of ASD (Tables 2 and 3).”

I rewrote the text according to the reviewer's instructions, in lines 144-145

“Having bedtimes of after 22:00 on weekdays and on weekends at infancy and at 1-3 years. of age were associated with the prevalence of ASD (Tables 2 and 3).”

 

Thank you for your kind comments and suggestions, I hope it has been corrected. correctly.