List of Papers By topics Author List
| Paper Info | Reviews | Meta-review | Author Feedback | Post-Rebuttal Meta-reviews |
Authors
Tingsong Xiao, Lu Zeng, Xiaoshuang Shi, Xiaofeng Zhu, Guorong Wu
Abstract
In this paper, we propose a dual-graph learning convolutional network (dGLCN) to achieve interpretable Alzheimer’s disease (AD) diagnosis, by jointly investigating subject graph learning and feature graph learning in the graph convolution network (GCN) framework. Specifically, we first construct two initial graphs to consider both the subject diversity and the feature diversity. We further fuse these two initial graphs into the GCN framework so that they can be iteratively updated (i.e., dual-graph learning) while conducting representation learning. As a result, the dGLCN achieves interpretability in both subjects and brain regions through the subject importance and the feature importance, and the generalizability by overcoming the issues, such as limited subjects and noisy subjects. Experimental results on the Alzheimer’s disease neuroimaging initiative (ADNI) datasets show that our dGLCN outperforms all comparison methods for binary classification. The codes of dGLCN are available on https://github.com/xiaotingsong/dGLCN.
Link to paper
DOI: https://link.springer.com/chapter/10.1007/978-3-031-16452-1_39
SharedIt: https://rdcu.be/cVVpU
Link to the code repository
https://github.com/xiaotingsong/dGLCN
Link to the dataset(s)
N/A
Reviews
Review #1
- Please describe the contribution of the paper
This paper proposed a dual-graph interpretable GCN to classify AD-NC and related MCI. And further, the proposed method could identify AD-related biomarkers.
- 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 paper makes a clear description of the dGLCN, which is a novel method to fuse the dual graphs from different group participants. The paper makes a detailed ablation analysis that make the method valid.
- 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.
Fig.2 is small. The version of tensorflow and GPU should be provided.
- 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 authors provided the code that make it is reproducible.
- 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 paper proposed a new method to classify AD-related groups and outperformed than other methods. It’s a good idea to provide more neuroimage information to make the interpretable method more substantial in Fig. 1.
- 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
7
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
This paper is well organized and give a new interpretable method, which is interesting and novel.
- Number of papers in your stack
1
- What is the ranking of this paper in your review stack?
5
- Reviewer confidence
Very confident
- [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
In this paper, dual-graph learning framework is proposed in the GCN context which has mainly three components including graph construction, dual-graph learning and graph fusion. Proposed framework is being utilized for the early AD diagnosis
- 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.
Problem is being posed in the context of dual representation learning. Graphs are being constructed at both the subjects and features level.
- 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.
Correlation is the metric being used for construction of subject graph and the feature graph. Is it optimal metric as inherent normality assumption is always there in this metric? Moreover, how to justify this statement statistically “correct correlations among the data are captured”?
- 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
Code is being provided for 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
Please define abbreviations everywhere in the text wherever they are being used first.
What is n and sigma in equation 1. Similarly, at others places variables are being used without defining them.
How the kNN being used for graph generation? More detail is need in this context.
Graphs are being updated after each and every layer. What does this signify? Practically, graph is being formed on an initial data which is than carried forward to all the layers in GCN.
Context of equ. 4 is not clear. Please explain. How the lambda_1 and lambda_2 parameters are being tuned?
How the size of architecture is being defined? Will A’ and S’ actually correspond to identifying interoperability?
Did the model convergence? It seems very less samples being considered in each class. Comparison on other public dataset with large samples is required in order to check generalization of framework.
- 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?
Overall the paper is well written and easy to follow. My one major concern is lack of enough samples while training and testing and use of conventional correlation metric for generating graphs.
- Number of papers in your stack
4
- What is the ranking of this paper in your review stack?
2
- Reviewer confidence
Very confident
- [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 article proposes to derive interpretable deep learning classifier (for Alzheimer’s disease) by jointly investigating subject and feature (structural ROIs) diversity. The model relies on jointly performing subject graph learning and feature graph learning within a graph convolutional network. The proposed method is more accurate than several baselines when applied to the cortical scores extracted from MRIs acquired by the Alzheimer’s disease neuroimaging initiative (ADNI) datasets.
- 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.
While the idea of accounting for noise in the labelling of subjects and brain ROI measurements is not new (see for example Adeli et al PAMI 2020 https://ieeexplore.ieee.org/document/8653343), doing so within a deep learning framework is intriguing. The authors nicely develop and carefully explain the model. The experiments are thorough especially for a conference paper. They consist of t-tests performed on the accuracy scores derived from repeated cross-validation applied to several classical machine learning approaches and more advanced deep learning models, and contain an ablation and parameter sensitivity study. Also the selection of ROI measures based on multiple trials is nice.
- 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.
The corresponding critics is minor in nature. Given that the authors include a sparse learning approach in their comparison, they should consider including the approach by Adeli et al. The description of the data set is too terse - it should at least include sex. Given that the topic is interpretability, the authors should at least discuss how confounders, such as sex, could influence their findings.
- 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
manuscript should specify which ADNI Data release was used for analysis
- 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
see above
- 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
7
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
see above
- Number of papers in your stack
4
- 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
N/A
- [Post rebuttal] Please justify your decision
N/A
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.
- The article proposes to derive interpretable deep learning classifier (for Alzheimer’s disease) by jointly investigating subject and feature (structural ROIs) diversity
- define abbreviations used in the paper
- 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).
3
Author Feedback
Many thanks to the reviewers and meta-reviewers for their valuable comments. We will address these issues in the final version.