List of Papers Browse by Subject Areas Author List
Abstract
Understanding the organization of human brain networks has become a central focus in neuroscience, particularly in the study of functional connectivity, which plays a crucial role in diagnosing neurological disorders. Advances in functional magnetic resonance imaging and machine learning techniques have significantly improved brain network analysis. However, traditional machine learning approaches struggle to capture the complex relationships between brain regions, while deep learning methods, particularly Transformer-based models, face computational challenges due to their quadratic complexity in long-sequence modeling. To address these limitations, we propose a Core-Periphery State-Space Model (CP-SSM), an innovative framework for functional connectome classification. Specifically, we introduce Mamba, a selective state-space model with linear complexity, to effectively capture long-range dependencies in functional brain networks. Furthermore, inspired by the core-periphery (CP) organization, a fundamental characteristic of brain networks that enhances efficient information transmission, we design CP-MoE, a CP-guided Mixture-of-Experts that improves the representation learning of brain connectivity patterns. We evaluate CP-SSM on two benchmark fMRI datasets: ABIDE and ADNI. Experimental results demonstrate that CP-SSM surpasses Transformer-based models in classification performance while significantly reducing computational complexity. These findings highlight the effectiveness and efficiency of CP-SSM in modeling brain functional connectivity, offering a promising direction for neuroimaging-based neurological disease diagnosis.
Our code is available at https://github.com/m1nhengChen/cpssm
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2025/paper/1080_paper.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to the Code Repository
https://github.com/m1nhengChen/cpssm
Link to the Dataset(s)
N/A
BibTex
@InProceedings{CheMin_CorePeriphery_MICCAI2025,
author = { Chen, Minheng and Yu, Xiaowei and Zhang, Jing and Chen, Tong and Cao, Chao and Zhuang, Yan and Lyu, Yanjun and Zhang, Lu and Liu, Tianming and Zhu, Dajiang},
title = { { Core-Periphery Principle Guided State Space Model for Functional Connectome Classification } },
booktitle = {proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2025},
year = {2025},
publisher = {Springer Nature Switzerland},
volume = {LNCS 15971},
month = {September},
page = {237 -- 247}
}
Reviews
Review #1
- Please describe the contribution of the paper
The main contribution of this paper lies in its proposed integration of a Core-Periphery State-Space Model (CP-SSM) with the Mamba architecture for neurological disorder classification. Specifically, the authors claim to enhance computational efficiency by replacing conventional transformer attention mechanisms with the Mamba module, which purportedly reduces quadratic complexity to linear scale while maintaining pattern capture capabilities. Furthermore, they introduce a neuroscience-inspired CP-SSM component after the Mamba block, arguing that its dual-state structure (core vs. periphery dynamics) better mimics the hierarchical information processing and connectivity patterns observed in biological neural systems. The methodological novelty is framed as bridging neuroimaging domain knowledge with state-space model adaptations, though the technical implementation appears incremental relative to existing Mamba variants and standard SSM literature. The authors validate their proposed framework on two widely recognized neuroimaging datasets: ABIDE and ADNI.
- Please list the major strengths of the paper: you should highlight 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.
Novel Integration of Core-Periphery Principles with MoE:While the core-periphery (CP) principle is not new in neuroscience, its explicit integration into a Mixture-of-Experts (MoE) framework (CP-MoE) represents a structured, biologically inspired approach to expert selection. This design leverages the brain’s intrinsic organizational properties (core nodes for efficient processing, periphery for integration) to guide model architecture, enhancing interpretability and alignment with neurobiological principles. Though not fully novel, this interdisciplinary application bridges computational modeling and neuroscience theory in a targeted manner. Efficient Long-Sequence Modeling with Mamba:The adoption of Mamba, a selective state-space model (SSM), addresses the quadratic complexity bottleneck of Transformers by achieving linear computational complexity. While SSMs like Mamba are not novel in themselves, their application to functional connectome classification is a fresh contribution.
- Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
Clarity:The authors claim that existing methods “often overlook the intrinsic characteristics of brain function during computational modeling, resulting in suboptimal performance in brain network analysis” as the rationale for introducing their core-periphery principle-guided state-space model. While this argument is conceptually plausible, the statement lacks explicit references to prior literature, weakening its credibility. Novelty: The statement “this is the first application of Mamba and MoE in functional brain network analysis” in the Conclusion requires refinement, as the integration of Mamba and Mixture-of-Experts (MoE) has been extensively studied and applied in the domain of Large Language Models (LLMs) prior to this work. While the adaptation to functional brain networks demonstrates domain-specific innovation, the framing should avoid overstating architectural novelty and instead emphasize its pioneering role in neuroscientific research. Significance: This study focuses primarily on accuracy, with little exploration of medical insights. This raises questions about the overall impact of the work: how do this work contribute to medical understanding or practice? Increasing accuracy alone is less meaningful in medical imaging research. Experiments: A notable concern arises from the reported performance of the BrainNetTF baseline on the ADNI dataset, where the accuracy (ACC = 70.3±4.7%) is stated to exceed both sensitivity (SEN = 69.7±7.6%) and specificity (SPE = 69.9±7.7%). The reported ACC exceeding both metrics raises concerns about potential inconsistencies in calculation, metric definitions, or statistical reporting.
- 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 has provided an anonymized link to the source code, dataset, or any other dependencies.
- Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2025/en/REVIEWER-GUIDELINES.html
N/A
- 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.
(3) Weak Reject — could be rejected, dependent on rebuttal
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
See the major weaknesses of the paper.
- Reviewer confidence
Confident but not absolutely certain (3)
- [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.
N/A
- [Post rebuttal] Please justify your final decision from above.
N/A
Review #2
- Please describe the contribution of the paper
The paper introduces a Core-Periphery State-Space Model (CP-SSM) for functional connectome classification, utilizing Mamba for linear complexity and CP-MoE for enhanced representation. It outperforms Transformer-based models in classifying neurological diseases like ASD and MCI, demonstrating improved efficiency and effectiveness.
- Please list the major strengths of the paper: you should highlight 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.
- Improved Performance Over Transformers: The CP-SSM surpasses Transformer-based models in classification performance on benchmark fMRI datasets, such as ABIDE and ADNI, while also significantly reducing computational complexity.
- Linear Complexity with Mamba: The use of Mamba, a selective state-space model, allows the CP-SSM to effectively capture long-range dependencies in functional brain networks with linear computational complexity, addressing the quadratic complexity issue faced by Transformer models.
- Core-Periphery Organization: By leveraging the core-periphery (CP) organization, which is a fundamental characteristic of brain networks, the CP-MoE model enhances the representation learning of brain connectivity patterns, improving the efficiency of information transmission and communication.
- Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
- The model heavily relies on the core-periphery organization as a guiding principle. This assumption might not hold true for all types of brain networks or across different populations, potentially limiting the model’s applicability.
- Section 2.3 needs clearer description, especially about the CP graph construction, what is the node feature for graph? What is the f_i(x), the expert module?
- From fig.2(b), top-4 expert selection seems work best on the ABIDE dataset, why finally using top-2 selection?
- The model’s performance can be influenced by the hyperparameter node rate r, different datasets have a different node rate, leading to limited generalizability across datasets.
- 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 has provided an anonymized link to the source code, dataset, or any other dependencies.
- Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2025/en/REVIEWER-GUIDELINES.html
N/A
- 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.
(4) Weak Accept — could be accepted, dependent on rebuttal
- 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 the design of Core-Periphery Principle is interesting
- Reviewer confidence
Very confident (4)
- [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.
N/A
- [Post rebuttal] Please justify your final decision from above.
N/A
Review #3
- Please describe the contribution of the paper
This paper proposes a core-periphery enhanced Mamba for functional connectome classification. The Mamba net is used to extract long-term dependencies in functional brain networks with linear complexity. A core-periphery-guided MoE is used to improve the learning of brain connectivity patterns. This paper also carried out comparative experiments on two datasets for evaluation.
- Please list the major strengths of the paper: you should highlight 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.
- Interesting work that combines Mamba with MoE for functional network classification.
- It is a well-written paper that is easy to follow.
- Experimental results showed that the proposed method can improve the performance compared with strong baselines.
- Please list the major weaknesses of the paper. Please provide details: for instance, if you state that a formulation, way of using data, demonstration of clinical feasibility, or application is not novel, then you must provide specific references to prior work.
- The experimental analyses lack depth. For example, in section 3.2, the authors simply stated the observation of performance advantage over baselines and did not explain the insight behind the results.
- The complexity and running time should be analyzed to show the advantages compared with Transformers.
- Please rate the clarity and organization of this paper
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 has provided an anonymized link to the source code, dataset, or any other dependencies.
- Optional: If you have any additional comments to share with the authors, please provide them here. Please also refer to our Reviewer’s guide on what makes a good review and pay specific attention to the different assessment criteria for the different paper categories: https://conferences.miccai.org/2025/en/REVIEWER-GUIDELINES.html
N/A
- 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.
(5) Accept — should be accepted, independent of rebuttal
- Please justify your recommendation. What were the major factors that led you to your overall score for this paper?
See the strengths and weaknesses.
- Reviewer confidence
Very confident (4)
- [Post rebuttal] After reading the authors’ rebuttal, please state your final opinion of the paper.
N/A
- [Post rebuttal] Please justify your final decision from above.
N/A
Author Feedback
N/A
Meta-Review
Meta-review #1
- Your recommendation
Provisional Accept
- If your recommendation is “Provisional Reject”, then summarize the factors that went into this decision. In case you deviate from the reviewers’ recommendations, explain in detail the reasons why. You do not need to provide a justification for a recommendation of “Provisional Accept” or “Invite for Rebuttal”.
N/A