Paper Info Reviews Meta-Review Author Feedback Post-rebuttal Meta-Reviews Back to top
List of Papers By topics Author List
Paper Info Reviews Meta-review Author Feedback Post-Rebuttal Meta-reviews

Authors

Antoine Grigis, Chloé Gomez, Vincent Frouin, Lynn Uhrig, Béchir Jarraya

Abstract

A major challenge in medicine is the rehabilitation of brain-injured patients with poor neurological outcome who experience chronic impairment of consciousness, termed minimally conscious state or vegetative state. Resting-state functional Magnetic Resonance Imaging (rs-fMRI) holds the promise of easy-to-acquire and wide-spectrum biomarkers. Previous rs-fMRI studies in monkeys and humans have highlighted that different consciousness levels are characterized by the relative prevalence of different functional connectivity patterns - also referred to as brain states - which conform closely the underlying structural connectivity. Results suggest that changes in consciousness lead to changes in connectivity patterns, not just the co-activation strength between regions, but also at the level of entire networks. In this work, a four-stage framework is proposed to identify interpretable spatial signature of consciousness, by i) defining brain regions of interest (ROIs) from atlases, ii) filtering and extracting the time series associated with these ROIs, iii) recovering disjoint networks and associated connectivities, and iv) performing pairwise non-parametric tests between network activities grouped by acquisition conditions. Our approach yields tailored networks, spatially consistent and symmetric. They will be helpful to study spontaneous recovery from disorders of consciousness known to be accompanied by a functional restoration of several networks.

Link to paper

DOI: https://link.springer.com/chapter/10.1007/978-3-031-16431-6_25

SharedIt: https://rdcu.be/cVD47

Link to the code repository

N/A

Link to the dataset(s)

N/A

Reviews

Review #1

  • Please describe the contribution of the paper

    The Modular Hierarchical Analysis (MHA) linear latent variable model was used to differentiate the various conditions of consciousness using resting-state fMRI. The statistical analysis showed the signature of consciousness from 5 monkey data.

  • Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

    The Modular Hierarchical Analysis (MHA) linear latent variable model was used to differentiate the various conditions of consciousness using resting-state fMRI. The statistical analysis showed the signature of consciousness from 5 monkey data.

  • Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.

    Authors used the Modular Hierarchical Analysis (MHA) linear latent variable model, which was already published by Monti et al. (2020), to uncover disjoint networks and associated activities for all subjects. For training, the MHA model is fitted on the 4 monkey dataset, and the log-likelihood is computed for 1 monkey validation dataset. Due to the limited number of subjects (only 5 subjects), the MHA model could be biased to the small number of subjects, and the uncover networks could be changed according to the sample size. Authors could describe how to interpret the difference of the disjoint networks according to the optimal K, and according to the different atlases (CoCMac, DictLearn, and CIVMR). The variation of functional patterns and associated activities across runs within same subject also could be examined for clarity.

  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

    Not Applicable

  • Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html

    Authors performed the experiments using more data. The variation of functional patterns and associated activities across runs within same subject also could be examined for clarity.

  • Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making

    5

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    limited sample size

  • Number of papers in your stack

    4

  • What is the ranking of this paper in your review stack?

    3

  • Reviewer confidence

    Confident but not absolutely certain

  • [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

    Not Answered

  • [Post rebuttal] Please justify your decision

    Not Answered



Review #2

  • Please describe the contribution of the paper

    This paper proposed a four-stage strategy and used a constraint linear latent variable model to define the spatial pattern of different consciousness states.

  • Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

    This work adopted a constraint linear latent varialble model to identify the spatial signatures of consciousness, which provides a novel application of the existing method and probably holds potential in clinical applications.

  • Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.
    1. The paper only adopted an existing method without any improvement and the references listed in the paper are not recent literatures, which weakens the novelty of this paper.
      Varol, A., Salzmann, M., Fua, P., & Urtasun, R. (2012, June). A constrained latent variable model. In 2012 IEEE conference on computer vision and pattern recognition (pp. 2248-2255).
    2. The adoption of three atlases in this work is not well explained and ambiguous.
    3. While the method can reveal the differences in network activities among different consciouness states, the inferred spatial patterns and dfifferent network activaties are group-wise, which can not be used, at least in current work, for subject-level recognition and thus limits the clinical application of proposed strategy.
  • Please rate the clarity and organization of this paper

    Good

  • Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

    The paper provides reasonable reproducibility.

  • Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html
    1. The paper only adopted an existing method without any improvement and the references listed in the paper are not recent literatures, which weakens the novelty of this paper.
      Varol, A., Salzmann, M., Fua, P., & Urtasun, R. (2012, June). A constrained latent variable model. In 2012 IEEE conference on computer vision and pattern recognition (pp. 2248-2255).
    2. The adoption of three atlases in this work is not well explained and ambiguous.
    3. While the method can reveal the differences in network activities among different consciouness states, the inferred spatial patterns and dfifferent network activaties are group-wise, which can not be used, at least in current work, for subject-level recognition and thus limits the clinical application of proposed strategy.
  • Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making

    3

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    The novelty and application value in clinical corhort are the major factors.

  • Number of papers in your stack

    4

  • What is the ranking of this paper in your review stack?

    3

  • Reviewer confidence

    Very confident

  • [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

    Not Answered

  • [Post rebuttal] Please justify your decision

    Not Answered



Review #4

  • Please describe the contribution of the paper

    This paper introduces a novel method for states of consciousness by understanding ROI connectivity patterns, proposing a new spatial biomarker for determining levels of consciousness.

  • Please list the main strengths of the paper; you should write about a novel formulation, an original way to use data, demonstration of clinical feasibility, a novel application, a particularly strong evaluation, or anything else that is a strong aspect of this work. Please provide details, for instance, if a method is novel, explain what aspect is novel and why this is interesting.

    1. The authors have a clear scientific/ medical understanding of the problem. This shows that the authors weren’t just applying cool technology to a common problem, but chose a framework that was backed by having a clinical understanding of consciousness.

    2. All sections of the paper, including the introduction which introduces the problem in a very detailed matter, were very clear and thorough

    3. The statistical analysis seems robust, in that data was tested for normality and after discovering it was non-normally distributed, switched to using non-parametric analysis.

    4. All graphs are easily interpretable.

  • Please list the main weaknesses of the paper. Please provide details, for instance, if you think a method is not novel, explain why and provide a reference to prior work.

    N/A

  • Please rate the clarity and organization of this paper

    Excellent

  • Please comment on the reproducibility of the paper. Note, that authors have filled out a reproducibility checklist upon submission. Please be aware that authors are not required to meet all criteria on the checklist - for instance, providing code and data is a plus, but not a requirement for acceptance

    The method is reproducible, although the data is not since it was acquired using a custom surface coil (although this is perfectly fine). The atlases used are also readily available, as they are open source atlases. Each section of the 4-step framework is thoroughly explained with all models written out.

  • Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review: https://conferences.miccai.org/2022/en/REVIEWER-GUIDELINES.html

    This was a solid paper that was very thorough, so there aren’t too many changes to be made.

    Methods

    1. In the ‘Atlases learned from the data’ section of the ‘ROIs definition’ section, there was mention of using ‘nilearn’. This should be referenced by (Python), along with the version.

    2. You should write out the definition of the acronym CIVMR before using it. I believe it’s (Center for In Vivo Microscopy), however, I am not sure where the ‘R’ comes from. This is based on the authors reference #8.

    3. A suggested word change: Maybe instead of using the word “perfect” in the last sentence of the ‘Discussion’ section, use the phrase “best tool thus far”.

  • Rate the paper on a scale of 1-8, 8 being the strongest (8-5: accept; 4-1: reject). Spreading the score helps create a distribution for decision-making

    8

  • Please justify your recommendation. What were the major factors that led you to your overall score for this paper?

    The paper introduces a novel framework for understanding connectivity patterns in different levels of consciousness. It does a great job of thoroughly explaining the problem and explaining the approach taken to solve the problem. The method is reproducible.

  • Number of papers in your stack

    6

  • What is the ranking of this paper in your review stack?

    1

  • Reviewer confidence

    Very confident

  • [Post rebuttal] After reading the author’s rebuttal, state your overall opinion of the paper if it has been changed

    Not Answered

  • [Post rebuttal] Please justify your decision

    Not Answered



Primary Meta-Review

  • Please provide your assessment of this work, taking into account all reviews. Summarize the key strengths and weaknesses of the paper and justify your recommendation. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. In case of an invitation for rebuttal, clarify which points are important to address in the rebuttal.

    This paper introduces a novel method for states of consciousness by understanding ROI connectivity patterns. The paper is well written, and tackles an important problem. Methodological novelty is verly limited and novelty lies mainly in the application. Some explanations as mentioned by reviewer 2 need to be added,.

  • What is the ranking of this paper in your stack? Use a number between 1 (best paper in your stack) and n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).

    8



Author Feedback

We thank the 3 reviewers for their insightful comments on the paper.

We agree with reviewers #1 and #2 that the methodology is not new, and the significant novelty lies in its successful application to the characterization of signatures of consciousness, a fundamental question cutting across medical imaging, cognitive neurosciences, and biomedicine. However, we do not understand the reviewer #2 considerations on the methodology, as it is not the claimed novelty of the paper.

We did reference recent neurosciences and methodological works. The applied MHA model is described by Monti et al. 2020 (and not by Varol et al. 2012, as stated by reviewer #2, which we did not refer to in the paper). While Varol and Monti both propose latent variable models (a common practice), Monti introduces specific constraints to map the resting-state fMRI signal to tailored networks depicting functional relationships between brain regions. The resulting model increases generalization capabilities, interpretability, and explainability. Unlike reviewer #2, we think the MHA model opens unique perspectives when searching for new disorders of consciousness biomarkers. Several studies associated different consciousness states with the activation of distinctive cortical areas. However, these results remain sparse. As far as we know, this is the first work to investigate brain-wide mechanisms in an unsupervised manner using the non-human primate loss of consciousness model. One of the discovered networks also supports the Global Neural Workspace theory promoting the brain network activities as innovative signatures of consciousness.

We agree with reviewer #1 that a small dataset size could bias the MHA model, resulting in potentially unstable networks. We recorded resting-state fMRI from 5 monkeys across states of consciousness, using distinct anesthetic agents. These are highly challenging experiments, and a sample size of N=5 is acceptable in the field of monkey fMRI. We further scanned monkeys multiple times over different conditions to generate a dataset with 156 samples (one sample = 1 monkey under 1 condition). Fair leave-one-subject-out cross-validation holds out all the data of a monkey during training. The risk of overfitting exists, but the results match well-established theories suggesting that the induced bias (if it exists) is reasonable.

Reviewer #2 questions the motivations under the adoption of three atlases. Working at the voxel level is computationally and statistically intractable as it involves modeling billions of connections. Standard approaches average signals on regions assuming temporal consistency within regions. This averaging should be carefully considered and results in the atlas selection problem described in section 2.1. Ultimately, the choice of the atlas depends on the scientific question that is under investigation. As suggested by reviewer #1, interpreting cross-atlas network differences is another interesting question. The atlases do not necessarily share the same spatial support nor integrate the same information or strategy during their development, as described in section 2.1. In this work, we pair the cross-atlas networks using a similarity measure described in section 2.3. The first paired networks share similar spatial locations, as depicted in section 3.3 and Fig. 4.

Reviewer #2 claims that the inferred network activities are group-wise and can not be used for subject-level recognition, thus limiting the clinical applications. However, as explained in sections 2.3 and 2.4, the inferred factors W describes reproducible networks across the entire population, while associated brain network activities are subject-specific. In other words, all subjects share the same loading matrix W, and the brain network activities allow subject-level recognition. On top of that, the model generalizes to unseen individuals, referred to as data not employed during the unsupervised learning of the latent variables.



Post-rebuttal Meta-Reviews

Meta-review # 1 (Primary)

  • Please provide your assessment of the paper taking all information into account, including rebuttal. Highlight the key strengths and weaknesses of the paper, clarify how you reconciled contrasting review comments and scores, indicate if concerns were successfully addressed in the rebuttal, and provide a clear justification of your decision. If you disagree with some of the (meta)reviewer statements, you can indicate so in your meta-review. Please make sure that the authors, program chairs, and the public can understand the reason for your decision.

    The reviewers have agreed that the application presented is interesting. Despite a limited methodological innovation, this paper might present an interesting discussion point at the conference.

  • After you have reviewed the rebuttal, please provide your final rating based on all reviews and the authors’ rebuttal.

    Accept

  • What is the rank of this paper among all your rebuttal papers? Use a number between 1/n (best paper in your stack) and n/n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).

    5



Meta-review #2

  • Please provide your assessment of the paper taking all information into account, including rebuttal. Highlight the key strengths and weaknesses of the paper, clarify how you reconciled contrasting review comments and scores, indicate if concerns were successfully addressed in the rebuttal, and provide a clear justification of your decision. If you disagree with some of the (meta)reviewer statements, you can indicate so in your meta-review. Please make sure that the authors, program chairs, and the public can understand the reason for your decision.

    The paper is original and interesting. Although the methodology per se is not new, the application to characterize signature of consciousness it is. A weakness is the limited number of subjects (although given the field is acceptable). The rebuttal satisfactorily addresses the comments by the reviewers.

  • After you have reviewed the rebuttal, please provide your final rating based on all reviews and the authors’ rebuttal.

    Accept

  • What is the rank of this paper among all your rebuttal papers? Use a number between 1/n (best paper in your stack) and n/n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).

    NR



Meta-review #3

  • Please provide your assessment of the paper taking all information into account, including rebuttal. Highlight the key strengths and weaknesses of the paper, clarify how you reconciled contrasting review comments and scores, indicate if concerns were successfully addressed in the rebuttal, and provide a clear justification of your decision. If you disagree with some of the (meta)reviewer statements, you can indicate so in your meta-review. Please make sure that the authors, program chairs, and the public can understand the reason for your decision.

    The strength is the neuroscience application is interesting to understand consciousness. The weakness is the methodology has limited innovation. The Introduction Section could run a literature research and better summarize the current status. Acceptance is recommended.

  • After you have reviewed the rebuttal, please provide your final rating based on all reviews and the authors’ rebuttal.

    Accept

  • What is the rank of this paper among all your rebuttal papers? Use a number between 1/n (best paper in your stack) and n/n (worst paper in your stack of n papers). If this paper is among the bottom 30% of your stack, feel free to use NR (not ranked).

    9



back to top