Gaze Dispersion During a Sustained-Fixation Task as a Proxy of Visual Attention in Children with ADHD

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors performed a rather simple oculomotor experiment addressing the fixational eye movements of children with attention deficit hyperactivity disorder (ADHD) and healthy control children. They performed a very simple analysis of the oculomotor data (taking only spatial values into account and leaving any temporal values out) and found very interesting differences between the two groups. In order to increase the strength of the argument that there are differences in fixational eye movements in ADHD children and control, the authors might include the following comments in their revision of the manuscript:

 

i) analysis of the different fixation targets as well as gap period

As far as I understand the presented data, the authors computed the BCEA for the complete measurement of 11 seconds for each subject. However, since three different fixation targets were shown for three seconds each, I strongly recommend to analyze each segment independently. In addition, two gap periods with one second each were shown. So, in total, the authors should present 5 BCEA-values and performed the appropriate statistical testing (i.e. 2-factorial ANOVA with factors group and stimulus condition).

 

ii) original data should be shown

The computation of the BCEA is not too complicated, and was already described in an earlier paper (Crossland et al 2002). However, I would also strongly suggest to show original data as scatter plot together with the resulting BCEA for a typical ADHD child and control child.

 

iii) important reference is missing

I would like to remind the authors to refer their data to:

Thaler, L., Schutz, A. C., Goodale, M. A., & Gegenfurtner, K. R. (2013). What is the best fixation target? The effect of target shape on stability of fixational eye movements. Vision Res, 76, 31-42.

 

iv) weakness of data

Fixational eye movements consists in micro-saccades, drift, and tremor. The quality of the obtained oculomotor data does not allow to perform a detailed analysis of the oculomotor data, i.e. to detect micro-saccades. I would like to remind the authors that micro-saccades represent a proxy for the direction of attention (see for instance: Hafed, Z. M., Lovejoy, L. P., & Krauzlis, R. J. (2011). Modulation of Microsaccades in Monkey during a Covert Visual Attention Task. J Neurosci, 31(43), 15219-15230.)

I see that the authors can not change the quality of their data, especially the poor temporal resolution of the Eya Eye Tracker. But I suggest that they address this issue offensively at the end of the abstract as well as in the discussion.

In line 106, the authors state that sampling frequency was set to 120 Hz. In line 137, they state that gaze position was recorded every 10 ms. Obviously, the two statements do not fit, sampling frequency was either 120 Hz (every 8.3 ms) or 100 Hz (every 10 ms).

Author Response

The authors performed a rather simple oculomotor experiment addressing the fixational eye movements of children with attention deficit hyperactivity disorder (ADHD) and healthy control children. They performed a very simple analysis of the oculomotor data (taking only spatial values into account and leaving any temporal values out) and found very interesting differences between the two groups. In order to increase the strength of the argument that there are differences in fixational eye movements in ADHD children and control, the authors might include the following comments in their revision of the manuscript:

 

Comment i) analysis of the different fixation targets as well as gap period

As far as I understand the presented data, the authors computed the BCEA for the complete measurement of 11 seconds for each subject. However, since three different fixation targets were shown for three seconds each, I strongly recommend to analyze each segment independently. In addition, two gap periods with one second each were shown. So, in total, the authors should present 5 BCEA-values and performed the appropriate statistical testing (i.e. 2-factorial ANOVA with factors group and stimulus condition).

Response i: We sincerely thank the reviewer for this pertinent comment. However, we would like to clarify that the BCEA analysis was deliberately limited to the visual fixation periods, i.e., the three 3-second sequences during which a fixation stimulus was displayed on the screen. The gap periods (1 second of black screen between targets) were not included in the calculation because they contain no visual cues to fixate on and were not designed to elicit active fixation.

Each of the three targets used in our protocol (see Figure 1) is presented for 3 seconds, and the children are instructed to look at the center of the stimulus. The first target is a classic cross (visual size: 1°) with a clear intersection point in the center. This first presentation serves as a clear and explicit visual anchor to initialize attentional calibration and illustrate the position of the center of fixation.

The next two targets are deliberately designed without an intersection in the center:

• the second cross is interrupted in the center, leaving an empty space 1° in diameter,
• the third is a diagonal cross with spread arms, also creating an empty center of approximately 1°.

This methodological choice is based on a strong hypothesis: by removing the visible central intersection, it becomes more difficult to maintain a precise fixation, which makes it possible to highlight any attentional and oculomotor instabilities, particularly in children with ADHD. Thus, the protocol was designed not to compare target shapes with each other, but to generate a prolonged and demanding fixation condition in a weakly structured visual context.

Our objective in this preliminary study was to obtain a global, continuous, and robust index of gaze stability in a deliberately unstructured context. For this reason, data from the three fixation conditions were aggregated for the calculation of the BCEA. This methodological choice was clarified in the Methods section and discussed in the Limitations section of the manuscript.

Nevertheless, a complementary study is currently underway with a larger sample, in which the BCEA will be analyzed separately for each type of target. This approach will allow us to better characterize the impact of the visual properties of the stimulus on fixation stability and to identify the conditions most sensitive to the specific attentional difficulties of children with ADHD.

 

Comment ii) original data should be shown

The computation of the BCEA is not too complicated, and was already described in an earlier paper (Crossland et al 2002). However, I would also strongly suggest to show original data as scatter plot together with the resulting BCEA for a typical ADHD child and control child.

Response ii: Figure 2 was changed according to your suggestion

 

Comment iii) important reference is missing

I would like to remind the authors to refer their data to:

Thaler, L., Schutz, A. C., Goodale, M. A., & Gegenfurtner, K. R. (2013). What is the best fixation target? The effect of target shape on stability of fixational eye movements. Vision Res, 76, 31-42.

Response iii: We thank the reviewer for this important reference. We have now included and discussed the findings of Thaler et al. (2013) in the Introduction and in the Discussion to support the rationale for using different fixation target types in our paradigm. The reference has been added to the bibliography.

 

Comment iv) weakness of data

Fixational eye movements consists in micro-saccades, drift, and tremor. The quality of the obtained oculomotor data does not allow to perform a detailed analysis of the oculomotor data, i.e. to detect micro-saccades. I would like to remind the authors that micro-saccades represent a proxy for the direction of attention (see for instance: Hafed, Z. M., Lovejoy, L. P., & Krauzlis, R. J. (2011). Modulation of Microsaccades in Monkey during a Covert Visual Attention Task. J Neurosci, 31(43), 15219-15230.)

I see that the authors can not change the quality of their data, especially the poor temporal resolution of the Eya Eye Tracker. But I suggest that they address this issue offensively at the end of the abstract as well as in the discussion.

Response iv: We agree that our eye-tracking system (sampling at 120 Hz) does not allow for reliable detection of microsaccades or tremor, limiting the granularity of our oculomotor analysis. We have now:

• Added a clarifying sentence at the end of the Abstract:

“Due to the temporal limitations of our recording system, this analysis focuses on gaze dispersion rather than fine-grained microsaccadic activity.”

• Expanded the Limitations section in the Discussion to explicitly acknowledge this limitation, cite Hafed et al. (2011), and clarify that our measure reflects overall gaze instability, not specific fixational components such as microsaccades.

 

Comment v) Sampling frequency inconsistency

In line 106, the authors state that sampling frequency was set to 120 Hz. In line 137, they state that gaze position was recorded every 10 ms. Obviously, the two statements do not fit, sampling frequency was either 120 Hz (every 8.3 ms) or 100 Hz (every 10 ms).

Response v: Thank you for spotting this inconsistency. You are absolutely right. The correct sampling frequency of the eye tracker is 120 Hz, which corresponds to one data point every 8.3 ms. We have corrected the sentence accordingly in the revised manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

This paper introduces a more effective measure for describing fixation instability in children with ADHD compared to the commonly used metric of intrusion saccades. The Bivariate Contour Ellipse Area (BCEA) is frequently utilized in various microperimetric studies, but it is rarely applied to analyze variability in gaze positions within neurodevelopmental disorders. Previous research, including some conducted by the authors themselves, has already documented fixation instability in children with ADHD.  Hence, it is not surprising that the references include numerous self-citations showing variability in fixation positions and intrusive saccades. The authors' application of the BCEA is similar to its usage in various oculomotor studies.  Thus, the contribution of the authors is to apply a well-established technique for the analysis of fixational instability to a specific neurodevelopmental disorder.

The authors defined their aim as exploring visual attention in children with ADHD using eye tracking and to identify a characteristic related to their attentional control. Still, the study only shows that BCEA is an appropriate measure of fixational instability that allows for discriminating the average eye behavior of the children with ADHD from that of children with typical development. 

One way to demonstrate the advantage of this technique could have been by highlighting its greater sensitivity compared to the saccade number during fixation. Another contribution of the study could have been if they had used the parameters involved in BCEA (the correlation between the horizontal and vertical eye positions (related to the direction of maximal elongation of BCEA) and the spread in the horizontal and vertical directions) to demonstrate some specific characteristics of the children's groups.   

It would be helpful to visualize the distributions of the gaze positions for the children with typical development and with ADHD. Moreover, the results presented in Figure 2 imply greater variability in BCEA of the ADHD group for the left eye as compared to the right one. Additionally, it is unclear why the authors employed different conditions to explore the participants' fixation behavior, as they do not compare the results of these conditions. 

There are also some issues with the cited literature. For example, the paper numbered 15 could not be found, and the cited doi corresponds to an Editorial. It might be good to include in the reference some new papers with meta-analyses of the oculomotor behavior of children with ADHD,

 

 

Author Response

Comment 1: “The authors defined their aim as exploring visual attention in children with ADHD using eye tracking and to identify a characteristic related to their attentional control. Still, the study only shows that BCEA is an appropriate measure of fixational instability that allows for discriminating the average eye behavior of the children with ADHD from that of children with typical development.” 

Response 1: We thank the reviewer for this pertinent observation. It is true that our study remains exploratory and focuses primarily on evaluating BCEA as a quantitative indicator of gaze stability in a sustained fixation paradigm. Our intention was not to provide an exhaustive assessment of visual attention in all its dimensions, but rather to test the hypothesis that BCEA could be an objective, continuous, and sensitive marker of attentional fluctuations in ADHD, based on simple and passive eye movement behavior.
We clarified this objective in the revised version of the abstract and introduction, emphasizing that we are presenting here a first differentiating factor between groups, measured through a minimally constraining fixation task. The advantage of this approach lies in its ease of clinical implementation and the fact that it does not rely on a complex task or a voluntary behavioral response.
We also specified in the discussion that future work will aim to explore in greater detail the links between BCEA and other cognitive attention indicators, as well as to examine correlations with individual clinical profiles.

Comment 2 : “One way to demonstrate the advantage of this technique could have been by highlighting its greater sensitivity compared to the saccade number during fixation. Another contribution of the study could have been if they had used the parameters involved in BCEA (the correlation between the horizontal and vertical eye positions (related to the direction of maximal elongation of BCEA) and the spread in the horizontal and vertical directions) to demonstrate some specific characteristics of the children's groups.”  

 
Response 2: We fully agree that a direct comparison between BCEA and other traditional indicators, such as the number of intrusive saccades, would have further highlighted the sensitivity of this measure. However, in this preliminary study, our main objective was to explore the relevance of BCEA as a continuous and integrated measure of gaze dispersion in a standardized protocol. The number of saccades during the fixation task was not systematically extracted in this first analysis.

Comment 3 : “It would be helpful to visualize the distributions of the gaze positions for the children with typical development and with ADHD. Moreover, the results presented in Figure 2 imply greater variability in BCEA of the ADHD group for the left eye as compared to the right one. Additionally, it is unclear why the authors employed different conditions to explore the participants' fixation behavior, as they do not compare the results of these conditions.” 

Response 3: We thank the reviewer for this feedback. These comments overlap with some of the points raised by Reviewer 1, which we have taken into account in the revised version of the manuscript. In particular:
•    We have added a new Figure 2, showing point clouds of gaze positions, visually illustrating the dispersion for a typical child and a child with ADHD. This visualization is specifically intended to respond to the request for representation of the raw data.
•    We have also clarified in the Methods section that the three fixation targets were designed to generate a continuous visual context, but that the analysis was performed on all stimulation periods, without distinction between targets, for methodological reasons explained in the Discussion. A complementary study is underway to analyze these conditions separately.
•    Finally, the slight difference in variability between the left and right eyes in the ADHD group, clearly visible in the figure, was not statistically significant according to our ANOVA. We have clarified this point in the Results section. 

Comment 4 :  “There are also some issues with the cited literature. For example, the paper numbered 15 could not be found, and the cited doi corresponds to an Editorial. It might be good to include in the reference some new papers with meta-analyses of the oculomotor behavior of children with ADHD”

Response 4: Reference No. 15 has been corrected, as the original DOI did indeed refer to an editorial by mistake. We have replaced this citation with a valid and complete reference corresponding to the original article used to support our argument.
In addition, we have supplemented our bibliography with several recent meta-analyses on oculomotor behavior in ADHD, notably those by Chamorro et al. (2022) and Maron et al. (2021), in order to update and strengthen the scientific basis of our study.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I like to thank the authors for their careful revision as well as for their detailed responses to my issues. I do not have additional comments.

Author Response

Thank you for your positive comment.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have taken most of my recommendations and concerns into account. I am convinced that BCEA is a better measure for evaluating the fixation instability than the saccade number. However, the authors do not fully explore the advantages of this approach. For example, the calculation of BCEA provides additional measures of fixation instability, such as the deviation of the center of the fixation positions as defined by the BCEA from the fixation target (regardless of whether it is a space or a crossing). Moreover, an interesting result might be a comparison of the spread of eye positions in horizontal and vertical directions. This information is available for the estimation of BCEA and could easily be estimated by standard methods for variance comparison.

The authors claimed that Figure 2 is an appropriate representation of the dispersion for a typical child and a child with ADHD. It is not what I was suggested. The figure illustrates the variability of BCEA, whereas I suggested presenting the distribution of fixation positions on which the estimation is based. It would have represented another characteristic of the fixation instability - whether the deviations of positions in x- and y-directions are correlated.

In summary, I understand the advantages of BCEA for the analysis of eye movements and, in particular, for describing the fixation instability in children with ADHD. However, this approach is used in other studies of gaze stability and is not new. In my understanding, showing that BCEA can represent the differences in gaze stability between typical children and children with ADHD is not enough; it is an expected result. I cannot understand why the authors do not fully exploit the advantages and the information that calculating BCEA provides.

A final comment - having higher temporal resolution in eye recordings would not allow for discriminating the different types of eye movements if BCEA is used for their analysis.

 

Author Response

Comments 1: [However, the authors do not fully explore the advantages of this approach. For example, the calculation of BCEA provides additional measures of fixation instability, such as the deviation of the center of the fixation positions as defined by the BCEA from the fixation target (regardless of whether it is a space or a crossing). Moreover, an interesting result might be a comparison of the spread of eye positions in horizontal and vertical directions. This information is available for the estimation of BCEA and could easily be estimated by standard methods for variance comparison.]

Response 1: We sincerely thank the reviewer for their constructive comments, which have enabled us to improve our manuscript and better define the scope of our approach. We fully understand the comment regarding the underutilized potential of the BCEA calculation. Indeed, beyond the dispersion area, this measure allows other interesting parameters to be extracted, such as:

• the offset of the point cloud's center of gravity relative to the fixation target;
• the asymmetry of horizontal and vertical dispersions, allowing for the exploration of a possible directional bias;
• the orientation of the ellipse (eigenvalues of the covariance matrix) and its eccentricity.

However, in this exploratory study, our main objective was to propose a simple method that could be applied in a clinical context, particularly for professionals working with children with ADHD. That is why we deliberately limited the analysis to an aggregated and automated measurement of the BCEA, which is easy to interpret and integrate into our analysis tool. The more in-depth analyses you suggest are indeed very relevant, and we plan to incorporate them into a complementary study currently underway. The aim of this study is to compare the different types of fixation targets in greater detail by analyzing the directional and structural parameters of the dispersion ellipse to better characterize attentional profiles, if they exist.

Comments 2 : [The authors claimed that Figure 2 is an appropriate representation of the dispersion for a typical child and a child with ADHD. It is not what I was suggested. The figure illustrates the variability of BCEA, whereas I suggested presenting the distribution of fixation positions on which the estimation is based. It would have represented another characteristic of the fixation instability - whether the deviations of positions in x- and y-directions are correlated.]

Response 2:  Regarding data visualization, we have taken note of your comment. The software currently used only plots the fixation point cloud and automatically calculates the BCEA at 68% confidence. A typical example of visualization (from our software) is shown below (see attached file). This rendering shows the spatial dispersion of fixations for each eye, although no explicit ellipse is shown at this stage. Developments are underway to enable automatic plotting of ellipses, incorporating the parameters mentioned above.

Comments 3: [A final comment - having higher temporal resolution in eye recordings would not allow for discriminating the different types of eye movements if BCEA is used for their analysis.]

Response 3 : Regarding your comment on the nature of BCEA and its limitation in distinguishing fine components of eye movements (such as microsaccades, drifts, or tremors), we have modified the section on limitations in the manuscript to explicitly clarify this point. We now clearly state that BCEA is not a method intended to identify or classify micro-movements, even with a high sampling frequency, and that it is primarily a global index of fixation stability.

 

Author Response File: Author Response.pdf