Content of review 1, reviewed on April 18, 2022
This study introduces a novel software to detect genuine smiles, differentiating with fake smiles that can be used in Dentistry. The topic is interesting and clinically. However, there are some points to be clarified or revised in the current manuscript.
Abstract
The current abstract does not read well. Since explanation on the experimental set-up necessary to be explained is lacking, it is hard to understand the overall method of this study.
Background: Rationale behind this study is unclear. Please explain why the quantitative analysis of smiles is needed in dentistry.
Method: “The following action units were assessed: Action Unit 6 (AU6 - cheek raiser), Action Unit 12 (AU12 - lip corner puller), and Action Unit 25 (AU25 - lips apart).” This description is not reader-friendly. I would suggest explaining more in detail on this detection system here.
What is a definition of “an experienced coder”? Since it is described as a gold standard, more detailed information is needed.
How genuineness and intensity was assessed quantitatively here?
Results: “While watching the clip,…..” what is “the clip”?
Introduction
In the 2nd paragraph, the relationship between analysis of smiles and dentistry is written. However, since this can be rationale behind this study, I would suggest expanding this part more.
In the 4th paragraph, some limitations on the manual detection of smiles are listed. I understood its complexity, but could the exact method to identify a smile episode manually be explained? In my opinion, if there is a standardized way to decide an onset of a smile episode, it would not matter if you are experienced or not.
Materials and methods
To differentiate “Duchenne smiles” from social smiles, more comprehensive objective recording of a participant is needed, not only video recordings. As the authors mention that a genuine smile induces muscle activities in the zygomaticus major and the orbicularis oculi muscles, it would be effective to record EMG activities of these muscles, as well as EEG for assessment of the emotional status. In this way, genuine smiles can be detected in an objective way without any subjective biases.
Page 5, Line 21: It is difficult to understand what AUs exactly are. Are they a group of specific points selected from 68 facial landmarks, according to facial movements? Moreover, it says that AU6, for example, tracks the activity of the orbicularis oculi muscle, which requires a more detailed information.
Conclusions
It would be great, if the authors can provide readers with more concrete advantage of this software that can be useful in dentistry.
Source
© 2022 the Reviewer.
Content of review 2, reviewed on June 29, 2022
The revised manuscript improved. However, there are still some points I would like the authors to consider.
Regarding the revised abstract, the first sentence does not seem professional and like a description in a scientific article. The rationale written here is not convincing.
The authors should avoid explaining the method, centering a device or a technology, not patients.
“While watching the amusing videos, …….for around one-third of the time.” The last part of this sentence (around 1/3 of the time) is an unclear expression.
“The onset and cessation of smile episodes were identified by two examiners trained with FACS coding, serving as a gold standard.” Subjective assessment cannot be a gold standard.
I understood that EMG activity of the zygomaticus major and the orbicularis oculi could be used to identify the genuine “Duchenne” smile and should be called “a gold standard”. In addition, again, subjective assessment cannot be a gold standard. Therefore, the authors should reconsider how to report their findings obtained in this study. According to comments from the authors answered to my previous comment, there is no standardized method to define the onset of a smile. It seems very beneficial to use the auto-detection of a smile with the software introduced in this manuscript. Therefore, a method to be compared should be carefully selected.
Source
© 2022 the Reviewer.
Content of review 3, reviewed on August 21, 2022
All points I pointed out have been addressed satisfactorily in this revised manuscript. I do not have any further comments.
Source
© 2022 the Reviewer.
References
Hisham, M., Jr., K. R., Hamza, B., B., H. J., Mauro, F. 2022. Automated detection of smiles as discrete episodes. Journal of Oral Rehabilitation.
