Do We Really Need Message Passing in Brain Network Modeling?
Authors: Liang Yang, Yuwei Liu, Jiaming Zhuo, Di Jin, Chuan Wang, Zhen Wang, Xiaochun Cao
ICML 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive evaluations verify the superiority of the proposed BQN compared to the message-passing-based brain network modeling. Source code is available at https: //github.com/LYWJUN/BQN-demo. [...] 6. Evaluations 6.1. Experiments Setup Datasets. In the experiments, two real-world f MRI datasets are employed: as follows. [...] Baselines. Thirteen Baselines are compared in the experiments, which can be divided into two categories. [...] Metrics. To conduct comprehensive evaluations of performance, a combination of machine learning and medical diagnostic-specific metrics are employed, including Area Under the Receiver Operating Characteristic Curve (AUC), Accuracy (ACC), Sensitivity (SEN), and Specificity (SPE). [...] 6.2. Result Analysis Brain Disorder Disease Classification. The performance comparison between BQN and the baseline models on the ABIDE and ADNI datasets are presented in Tab. 1. |
| Researcher Affiliation | Academia | 1Hebei Province Key Laboratory of Big Data Calculation, School of Artificial Intelligence, Hebei University of Technology, Tianjin, China 2Tianjin Key Laboratory of Cognitive Computing and Application, College of Intelligence and Computing, Tianjin University, Tianjin, China 3School of Computer Science and Technology, Beijing Jiao Tong University, Beijing, China 4School of Artificial Intelligence, OPtics and Electro Nics (i OPEN), School of Cybersecurity, Northwestern Polytechnical University, Xi an, China 5School of Cyber Science and Technology, Shenzhen Campus of Sun Yat-sen University, Shenzhen, China. |
| Pseudocode | No | The paper describes iterative formulas and mathematical equations for the models (e.g., Eq. 1-4, 7-10, 14-15) and provides theoretical analysis, but it does not include a dedicated pseudocode or algorithm block. |
| Open Source Code | Yes | Source code is available at https: //github.com/LYWJUN/BQN-demo. |
| Open Datasets | Yes | Autism Brain Imaging Data Exchange (ABIDE): This dataset is primarily utilized to investigate brain functional connectivity variation and structural differences associated with Autism Spectrum Disorder. The preprocessed data version can be accessed from the official website1. [...] Alzheimer s Disease Neuroimaging Initiative (ADNI): ADNI is a widely utilized multimodal neuroimaging repository focused on Alzheimer s disease. The raw images can be obtained from ADNI official website 2. [...] In Appendix E: (1) Attention Deficit Hyperactivity Disorder (ADHD-200) dataset is a collaboration of 8 international imaging sites [...] (2) Parkinson s Progression Markers Initiative (PPMI) dataset aims to identify biological markers of Parkinson s risk, onset and progression. [...] For PPMI dataset, the node attributes and edge weights were preprocessed by authors in (Xu et al., 2023). |
| Dataset Splits | Yes | For all datasets, we employ random splits with the ratio 7:1:2 to get the training set, validation set and test set. [...] In Appendix E, Implementation Details: For the above two datasets, random partitions were used with a ratio of 7 : 1 : 2. |
| Hardware Specification | Yes | The experiments are performed on a Ge Force RTX3090 GPU. |
| Software Dependencies | No | The paper mentions using an 'adam optimizer' and 'Leaky ReLU' activation function but does not specify versions for any programming languages, libraries, or frameworks (e.g., Python, PyTorch, TensorFlow, CUDA). |
| Experiment Setup | Yes | Furthermore, the number of training epochs is set to 200 with batch size 16. An adam optimizer is adopted with initial learning rate of 10 4 and weight decay as 10 4 while training, target learning rate is from 10 5 to 10 4. The activation function we selected is Leaky ReLU. The number of layers, i.e., k is selected from 1, 2, 3, 4, 5 and the dropout rate is chosen from 0., 0.1, 0.2, 0.3. The results were averaged over 5 random runs. |