Abstract

A close relation between brain function and cortical folding has been demonstrated by macro-/micro- imaging, computational modeling, and genetics. Since gyri and sulci, two basic anatomical building blocks of cortical folding patterns, were suggested to bear different functional roles, a precise mapping from brain function to gyro-sulcal patterns can provide profound insights into both biological and artificial neural networks. However, there lacks a generic theory and effective computational model so far, due to the highly nonlinear relation between them, huge inter-individual variabilities and a sophisticated description of brain function regions/networks distribution as mosaics, such that spatial patterning of them has not been considered. To this end, as a preliminary effort, we adopted brain functional gradients derived from resting-state fMRI to embed the “gradual” change of functional connectivity patterns, and developed a novel attention mesh convolution model to predict cortical gyro-sulcal segmentation maps on individual brains. The convolution on mesh considers the spatial organization of functional gradients and folding patterns on a cortical sheet and the newly designed channel attention block enhances the interpretability of the contribution of different functional gradients to cortical folding prediction. Experiments show that the prediction performance via our model outperforms other state-of-the-art models. In addition, we found that the dominant functional gradients contribute less to folding prediction. On the activation maps of the last layer, some well-studied cortical landmarks are found on the borders of, rather than within, the highly activated regions. These results and findings suggest that a specifically designed artificial neural network can improve the precision of the mapping between brain functions and cortical folding patterns, and can provide valuable insight of brain anatomy-function relation for neuroscience.

Links to Paper and Supplementary Materials

Main Paper (Open Access Version): https://papers.miccai.org/miccai-2024/paper/2888_paper.pdf

SharedIt Link: pending

SpringerLink (DOI): pending

Supplementary Material: https://papers.miccai.org/miccai-2024/supp/2888_supp.pdf

Link to the Code Repository

N/A

Link to the Dataset(s)

N/A

BibTex

@InProceedings{Yan_Brain_MICCAI2024,
        author = { Yang, Li and He, Zhibin and Zhong, Tianyang and Li, Changhe and Zhu, Dajiang and Han, Junwei and Liu, Tianming and Zhang, Tuo},
        title = { { Brain Cortical Functional Gradients Predict Cortical Folding Patterns via Attention Mesh Convolution } },
        booktitle = {proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2024},
        year = {2024},
        publisher = {Springer Nature Switzerland},
        volume = {LNCS 15007},
        month = {October},
        page = {pending}
}


Reviews

Review #1

  • Please describe the contribution of the paper

    The paper proposes a model for segmenting a cortical surface mesh into gyri and sulci areas using a mesh annotated with functional gradients. They employ a U-Net mesh attention convolution that independently weights the importance of each gradient channel, learned during training. The approach is compared to state-of-the-art methods for this task, along with conducting ablation studies on different types of attention models and the number of functional gradient channels.

  • 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 paper presents a straightforward formulation of the problem, making it easy to follow. The experiments are well-designed and demonstrate improvements over state-of-the-art methods. Additionally, the paper includes ablation studies to validate the modelling decisions made.

  • 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.
    • In tasks like this, the ground-truth information (gyri and sulci labels) is often a pseudo-ground truth derived from another estimation method rather than being annotated by an expert. It’s unclear if this is the case in this paper. If so, providing a qualitative evaluation comparing where these approaches diverge or adhere closely to the pseudo-ground truth could help identify data bias.
    • Additionally, what are the advantages of using a model to predict these regions from functional gradients rather than from structural MRI images, as done in traditional neuroimaging pipelines like FreeSurfer?
  • Please rate the clarity and organization of this paper

    Very Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission does not mention open access to source code or data but provides a clear and detailed description of the algorithm to ensure reproducibility.

  • Do you have any additional comments regarding the paper’s reproducibility?

    N/A

  • Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review. Pay specific attention to the different assessment criteria for the different paper categories (MIC, CAI, Clinical Translation of Methodology, Health Equity): https://conferences.miccai.org/2024/en/REVIEWER-GUIDELINES.html

    The simple formulation proposed in the paper doesn’t impose any specific topology on the final segmentation, which may result in fragmented segmentation. It seems that adopting approaches to address this issue could enhance the model’s performance. It would be beneficial to hear from the authors about the reasons for not exploring these approaches.

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making

    Weak Accept — could be accepted, dependent on rebuttal (4)

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

    The paper explains the problem well and shows improvements over existing methods in its experiments. However, there are some questions about the ground-truth labels, the topology of the predicted segmentation, and what are the advantages of using functional gradients instead of structural MRI images. If the authors can clarify these points and discuss why their approach is better, the paper could be accepted with some reservations.

  • Reviewer confidence

    Somewhat confident (2)

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

    N/A

  • [Post rebuttal] Please justify your decision

    N/A



Review #2

  • Please describe the contribution of the paper

    Authors investigate the predictive ability of brain functional gradients on cortical folding patterns at the macro scale using an attention mesh convolution model. Their model considers the spatial organization of functional gradients and folding patterns, while previous methods only used independent functional signal from vertices. One run of experimental results by the proposed model outperformed state-of-the-art.

  • 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 proposed model performs the best on gyri/sulci segmentation.

    2. Ablation studies show the effectiveness of the proposed modules.

    3. Unpaired t-test shows a conclusion of the relationship between functional activation value with gyral peaks and sulcal pits.

  • 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. Although high-level idea in this paper is easy to understand, the description of method and Fig.1(a) is not clear to me. How’s the functional gradient on cortical vertex is calculated?

    2. The organization of text is not efficient. Some paragraph is so long that it is hard to follow. Like Discussiom.

    3. There is only 1 run of deep learning experiments. This weakens the conclusion based on the experiments.

  • Please rate the clarity and organization of this paper

    Satisfactory

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission does not mention open access to source code or data but provides a clear and detailed description of the algorithm to ensure reproducibility.

  • Do you have any additional comments regarding the paper’s reproducibility?

    N/A

  • Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review. Pay specific attention to the different assessment criteria for the different paper categories (MIC, CAI, Clinical Translation of Methodology, Health Equity): https://conferences.miccai.org/2024/en/REVIEWER-GUIDELINES.html

    See weakness.

  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making

    Weak Accept — could be accepted, dependent on rebuttal (4)

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

    Experimental results in this paper are intriguing, but paper writing can be improved.

  • Reviewer confidence

    Very confident (4)

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

    N/A

  • [Post rebuttal] Please justify your decision

    N/A



Review #3

  • Please describe the contribution of the paper

    The author designed a novel attention mesh convolution method based on U-net architecture to predict the gyro-sulcal segmentation map from brain functional gradient maps.

  • 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 combination of mesh convolution and channel attention block is novel.
    2. Provide visualiztions of segmentation results.
    3. The paper is well-written.
  • 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.

    In the paper, the model is based on UNet architecture. The author incorporate the attention mask convolution into UNet, therefore other SOTA varients of UNet or SOTA segmentation models should be compared in Table 1, such as:

    1. TransUNet—> https://github.com/Beckschen/TransUNet
    2. UNeXt —> https://github.com/jeya-maria-jose/UNeXt-pytorch
  • Please rate the clarity and organization of this paper

    Very Good

  • Please comment on the reproducibility of the paper. Please be aware that providing code and data is a plus, but not a requirement for acceptance.

    The submission does not mention open access to source code or data but provides a clear and detailed description of the algorithm to ensure reproducibility.

  • Do you have any additional comments regarding the paper’s reproducibility?

    N/A

  • Please provide detailed and constructive comments for the authors. Please also refer to our Reviewer’s guide on what makes a good review. Pay specific attention to the different assessment criteria for the different paper categories (MIC, CAI, Clinical Translation of Methodology, Health Equity): https://conferences.miccai.org/2024/en/REVIEWER-GUIDELINES.html

    In the paper, the model is based on UNet architecture. The author incorporate the attention mask convolution into UNet, therefore other SOTA varients of UNet or SOTA segmentation models should be compared in Table 1, such as:

    1. TransUNet—> https://github.com/Beckschen/TransUNet
    2. UNeXt —> https://github.com/jeya-maria-jose/UNeXt-pytorch
  • Rate the paper on a scale of 1-6, 6 being the strongest (6-4: accept; 3-1: reject). Please use the entire range of the distribution. Spreading the score helps create a distribution for decision-making

    Weak Accept — could be accepted, dependent on rebuttal (4)

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

    The proposed model is novel and achieve good performance in this Gyro-sulcus segmentation task. Need more SOTA segmentation models for comparison study.

  • Reviewer confidence

    Somewhat confident (2)

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

    N/A

  • [Post rebuttal] Please justify your decision

    N/A




Author Feedback

N/A




Meta-Review

Meta-review not available, early accepted paper.



back to top