Neurofeedback Effects on EEG Connectivity among Children with Reading Disorders: I. Coherence

Round 1

Reviewer 1 Report

Dear authors,

 

Overall, I suggest avoiding terms such as "we have/had." Try to replace them with informal speech. At the end of the introduction, I notice the wording "we expect." This kind of word has a strong meaning, almost as if the authors would do anything to have the results proposed. I'd suggest deleting the last paragraph of your introduction.

Introduction:

I'm concerned with the affirmation made in lines 262 to 264. The protocol is important and can influence the final results. Recent studies trying to find the influence of different methodologies showed that the number of sessions (1Domingos et al., 2021a) and even the noise (Domingos et al., 2021b)could contribute to different outcomes when applying NFB.

Please, complement this information or delete that paragraph.

1Domingos, C., Peralta, M., Prazeres, P., et al. Session Frequency Matters in Neurofeedback Training of Athletes. Appl Psychophysiol Biofeedback 46, 195–204 (2021). https://doi.org/10.1007/s10484-021-09505-3

2Domingos, C.; da Silva Caldeira, H.; Miranda, M.; Melício, F.; Rosa, A.C.; Pereira, J.G. The Influence of Noise in the Neurofeedback Training Sessions in Student Athletes. Int. J. Environ. Res. Public Health 2021, 18, 13223. https://doi.org/10.3390/ijerph182413223

Lines 299 to 303. I would transfer this information to the Materials and Methods section.

Materials and Methods:

Did you run a power sample statistic? Because you started with 204 children, and your final sample was 46. If you did not run a power sample, how can you say your sample is enough to prove your hypothesis?

Again, you use the word "expected."

Saying that the sample size is not very large and that a normal distribution is not guaranteed, so you had to use non-parametric statistics seems forced. It almost sounds like you used non-parametric statistics to find significant differences since parametric statistics will show no differences.

 

If you are using a non-parametric statistic, you need valid arguments (published in the literature) and to perform a normality test to be sure.

Author Response

The authors are very grateful to the editor and the reviewers of this study. We appreciate your respectful comments, which have been aimed at improving this manuscript.

Below we answer the questions and recommendations made by the Reviewer-1.

1.- Overall, I suggest avoiding terms such as "we have/had." Try to replace them with informal speech. At the end of the introduction, I notice the wording "we expect." This kind of word has a strong meaning, almost as if the authors would do anything to have the results proposed. I'd suggest deleting the last paragraph of your introduction.
A: We made some changes to the wording, eliminating the allusion to the first-person plural. Hypotheses related to the main objective were deleted at the end of the Introduction.

Introduction:
2.- I'm concerned with the affirmation made in lines 262 to 264.
"This reduction in the theta/alpha ratio in both groups could be attributed to various factors not specific to the operant conditioning process (closely linked to the NFB protocol) but that could be present in both treatment conditions (NFB and NFB-sham), such as expectancy [61] or placebo [62–64] effects. Furthermore, both conditions involve the use of metacognitive strategies, which have been shown to modify the EEG [65, 66]."
The protocol is important and can influence the final results. Recent studies trying to find the influence of different methodologies showed that the number of sessions (1Domingos et al., 2021a) and even the noise (Domingos et al., 2021b)could contribute to different outcomes when applying NFB.
Please, complement this information or delete that paragraph.
1Domingos, C., Peralta, M., Prazeres, P., et al. Session Frequency Matters in Neurofeedback Training of Athletes. Appl Psychophysiol Biofeedback 46, 195–204 (2021). https://doi.org/10.1007/s10484-021-09505-3
2Domingos, C.; da Silva Caldeira, H.; Miranda, M.; Melício, F.; Rosa, A.C.; Pereira, J.G. The Influence of Noise in the Neurofeedback Training Sessions in Student
Athletes. Int. J. Environ. Res. Public Health 2021, 18, 13223. https://doi.org/10.3390/ijerph182413223
A: The authors fully agree that the protocol is essential for many reasons. This is one of the main NFB research interests of our group. According to our way of operating, to apply NFB to an individual who has a certain pathology, he/she must exhibit an abnormality related to said pathology in his/her EEG. The protocol involves the frequency of the EEG to be modified, the leads (or source), the sensory modality of the reinforcer, the valence of the reinforcer (positive or negative), the number of sessions, the time of each session, etc. Of course, all this has a direct influence on the results. We are delighted to see that an effort is being made to determine the optimal number of sessions, perhaps modulated by factors that are difficult to control (age/plasticity, the severity of the pathology, etc.). In our statement, we do not deny this.
We mean that both the NFB and the Sham-NFB produce some changes (given by metacognitive strategies of self-observation and monitoring, or by expectation, etc.). In addition to these changes common to both treatments, NFB produces specific changes, which in our case are seen as an improvement in reading, together with normalization of EEG coherence patterns.

3.- Lines 299 to 303. I would transfer this information to the Materials and Methods section.
A: We have rewritten the rationale that conduces to propose the first stage of the study (RD vs. TD) in lines 301-319 of the new version. Transferring this to the Materials and Methods section was unnecessary because this was already explained there.

Materials and Methods:
4.- Did you run a power sample statistic? Because you started with 204 children, and your final sample was 46. If you did not run a power sample, how can you say your sample is enough to prove your hypothesis?
A: At the beginning of the Participants subsection of the Materials and Methods, we included our calculation of the necessary sample sizes for the first and the second stages in order to have an a=0.05 and Power (1-b)=0.8
According to these calculations, we needed 18 subjects per group for the first stage, and for the second stage, 4 subjects are enough.

5.- Again, you use the word "expected."
A: We deleted "expected" in the Method section.

6.- Saying that the sample size is not very large and that a normal distribution is not guaranteed, so you had to use non-parametric statistics seems forced. It almost sounds like you used non-parametric statistics to find significant differences since parametric statistics will show no differences.
If you are using a non-parametric statistic, you need valid arguments (published in the literature) and to perform a normality test to be sure.
A: Multivariate parametric statistical analyzes require that Multivariate Normality be satisfied. However, for Multivariate Normal data, marginal distribution and linear combinations should also be Normal; therefore, it is enough that a marginal distribution is not Normal, so this requirement is not fulfilled.
Because in all cases n<50, we performed the Kolmogorov-Smirnoff test to asses Normality. H0 (Normality) is rejected when p<0.05. Four excel files with the results of the Kolmogorov-Smirnoff test are attached. In summary:
COGNITIVE VARIABLES
None of the cognitive variables had a Normal distribution for either of the two groups, TD and RD, in the first stage. In the second stage, none of the cognitive variables had a
Normal distribution for either of the two groups, NFB and SHAM, for neither of the two conditions (before and after).
THETA/ALPHA RATIO
Theta/alpha ratio did not satisfy Normality for either of the two groups (NFB, SHAM) in any of the two conditions (BEFORE, AFTER).
COHERENCE VARIABLES
In the first stage, 39% of the coherence variables were not distributed as Normal in the TD group, and 60% of the coherence variables were not distributed as Normal in the RD group.
In the second stage, 63% of the coherence variables were not distributed as Normal in the NFB group before treatment, and 31% of the coherence variables were not distributed as Normal in the NFB group after treatment. In contrast, 13% of the coherence variables were not distributed as Normal in the SHAM group before treatment, and 7% of the coherence variables were not distributed as Normal in the NFB group after treatment.
This leads us to the fact that it is not appropriate to use parametric statistics. Therefore we used non-parametric statistics; in particular, we used a multivariate analysis of permutations (Galan et al., 1997). In addition, we wrote some of the advantages of using a permutation test due to the multivariate nature of our data.

Author Response File: Author Response.pdf

Reviewer 2 Report

Judging by all publications (this one and refs 45-47, 56, 60), the authors have developed a kind of research style, which limits the study significance by a small number of participants. In future research, I recommend to correct this style.

The authors should correct the reference list, since under the reference numbers 24 and 33 there is the same article.

 

Updated on 2.2 

1.     What is the main question addressed by the research?

The main question addressed by the research is to explore the changes in EEG coherence patterns in children with RD and EEG maturational lag treated with NFB using a protocol that reinforces a reduction in the theta/alpha ratio.

2.     Do you consider the topic original or relevant in the field? Does it
address a specific gap in the field?

The topic is original. It addresses a specific gap in the field.

3.     What does it add to the subject area compared with other published
material?

NFB can positively impact reading-related functions in the brain networks of children with RD.

4.     What specific improvements should the authors consider regarding the
methodology? What further controls should be considered?

In future studies it is recommended to use more representative groups of subjects.

5.     Are the conclusions consistent with the evidence and arguments presented
and do they address the main question posed?

Yes, the conclusions are consistent with the evidence and arguments presented.

6.     Are the references appropriate?

Yes, the references are appropriate, however, within the last 5 years there are only 12 from 95 references (13%).

7.     Please include any additional comments on the tables and figures.

Tables and figures are appropriate and clear.

Comments for author File: Comments.pdf

Author Response

The authors are very grateful to the editor and the reviewers of this study. We appreciate your respectful comments, which have been aimed at improving this manuscript.
Below we answer the questions and recommendations made by the Reviewer-2:

Judging by all publications (this one and refs 45-47, 56, 60), the authors have developed a kind of research style, which limits the study significance by a small number of participants. In future research, I recommend to correct this style.

A: We acknowledge that the studies carried out by our group limit the application of the results to a small number of individuals. So far, we have proceeded this way because our goal has been to deeply understand this small population. However, we are also interested in being able to generalize the findings to larger populations. We appreciate the reviewer's recommendation for future research.

The authors should correct the reference list, since under the reference numbers 24 and 33 there is the same article.
A: The reference list was corrected.

Round 2

Reviewer 1 Report

Nothing to add