Development of a Multiple RGB-D Sensor System for ADHD Screening and Improvement of Classification Performance Using Feature Selection Method

Round 1

Reviewer 1 Report

Interesting article. In biological sciences, we follow IMRAD method of presenation. Researach questions/objectives are included at the end of the Introduction part of the paper. Sections I-IV are mentioned, but paper doesnt have sections.

My specific comments for authors :

Comment 01: Need details of the expertise of the clinicians who gave feedback: age, years of experience in the speciality e.g., psychiatrist, neurologist, pediatrician, and pediatric neurologist. Experince of having diagnosed children with ADHD. Even clinical psychologists, and school psychologist recognize ADHD.

 

Comment 02:  ADHD is a clinical diagnosis. The overall approach to diagnosis may involve (1) a comprehensive interview with the child’s adult caregivers; (2) a mental status examination of the child; (3) a medical examination for general health and neurological status; (4) a cognitive assessment of ability and achievement; (5) use of ADHD-focused parent and teacher rating scales; and (6) school reports and other adjunctive evaluations if necessary (speech, language assessment, etc.)[Gualtieri CT, Johnson LG. ADHD: Is Objective Diagnosis Possible? Psychiatry (Edgmont). 2005 Nov;2(11):44-53. PMID: 21120096; PMCID: PMC2993524.]

Following are more popular screening tools: ADHD rating scale, Connor screening scal efor ADHD, Vanderbilt scale for ADHD.

However, computer based technology is trying to make it more objective. Please have a look at TOVA. T.O.V.A(Test of Variables of Attention) is a neuropsychological assessment that measures a person's attention while screening for attention deficit hyperactivity disorder.

Comment 03:  Need to consider overlapping conditions that also affect movements: behavioral or conduct problems, anxiety disorders, depression, autism spectrum disorder, Tourette syndrome.

 

Comment 04:  existing screening tools for ADHD have limitations. Better use Gold standard: Consensus Case Diagnosis arrived by two clinicians.structured Clinical interview based on DSM-5( American Psychiatric Association, 2013) by Proxy version(Parent- and teacher interviewed) are used. SCID or Kiddie Schedule of Affective Disorders and Schizophrenia for School-Aged Children Present and Lifetime Version (KSADS-PL) are used. Consensus Case diagnostic decision arrived by at least two experts is near to the gold standard.

 

Comment 05: Authors showed limitation of previous screening due to contact interface and not real life scenario. But however, I see there method is far awar from real-life situation that child is required to go through: school, while doing homework, at play area, in the personal spaces,  in the public places.

 

Comment 06: As per the authors : The second type of the “A robot-led game for the screening of ADHD“ is a 1-versus-187 3 (1:3) domino game, not analysed due to small sample size.  Giving description of this , and explaining in the figures dilutes the paper strenght. I suggest delete all things related to 1:3 domino game.

Author Response

Thanks for your comment. Please check the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments for author File: Comments.pdf

Author Response

Thanks for your comment. Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

None

Reviewer 2 Report

   Although the authors make significant improvements in the version, there are still some issues not being exactly clarified (particularly for items listed in Concern #2 of the previous review comments), however.  Therefore, the authors are suggested to deal with the following issues by providing a more detailed explanation to spotlight the contributions and findings of this paper.

 

1. Referring to the Concern #2.1: The relationships between Figure 15 (model used in the experiment) and that of the three experiments (or the corresponding RQs) mentioned in session 4 are still hard to see the clear linkage amongst them.

2. Referring to the Concern #2.2:   The authors provide almost no further revision or clarification in terms of the truly perdition targets / purposes of this paper (even though the Abstract has been revised).

3. Referring to the Concern #2.3: How A1 is formed? Is A1 alone itself strong enough for ensuring the model prediction power? Why do the other factors identified as Rank1 (as shown in Table 3) provide merely limited contribution in model prediction (i.e., why these factors are good enough regarded as Rank 1 candidates)?

4. Derives from Concern #2.3: The authors still provide no clues in terms of how the 15 features are grouped into four factors (with a validated process). As well, the reasonability of definitions of features coming after feature grouping should be explained. Meanwhile, since all the discussion around the whole paper is based on feature-level (rather factor-level), the authors are suggested to consider the necessity and the meaning of forming factors (particularly with the case of having the 4th factor as “others”)

5. Further minor suggestions: (1) be sure the research title clearly present the key contribution and research focus of this paper; (2) be sure the whole presentation structure is understandable (i.e., the linkage and casual relationships amongst paragrams in different sections are easily understood from the readers’ perspective); (3) be sure Figure 5 is necessary (i.e., if the algorithm shown in the Figure is critical, be sure detailed description exists in the paragraph; in contrast, if the context shown in the Figure is supplementary for this study, it is suggested to remove it).

Comments for author File: Comments.pdf

Author Response

Thank you for your suggestion. Please see the attachment

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

Comments for author File: Comments.pdf

Author Response

Thank you for the comments, please see the attachment.

Author Response File: Author Response.pdf

Round 4

Reviewer 2 Report

Accept in present form