EC-GAN: Inferring Brain Effective Connectivity via Generative Adversarial Networks
Authors: Jinduo Liu, Junzhong Ji, Guangxu Xun, Liuyi Yao, Mengdi Huai, Aidong Zhang4852-4859
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on simulated data show that EC-GAN can better infer effective connectivity compared to other state-of-the-art methods. The real-world experiments indicate that EC-GAN can provide a new and reliable perspective analyzing the effective connectivity of f MRI data. |
| Researcher Affiliation | Academia | 1Beijing Key Laboratory of Multimedia and Intelligent Software Technology, Beijing Artificial Intelligence Institute, Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China 2Department of Computer Science, University of Virginia, Charlottesville, Virginia 22904, USA 3Department of Computer Science and Engineering, SUNY at Buffalo, Buffalo, New York 14260, USA |
| Pseudocode | Yes | Algorithm 1 shows the full details of the proposed EC-GAN. |
| Open Source Code | No | The paper does not provide an explicit statement or link indicating that the source code for EC-GAN is openly available. It only mentions that |
| Open Datasets | Yes | The benchmark simulation datasets we used are generated by (Smith et al. 2011) and (Sanchez-Romero et al. 2019a), which are widely used for detecting methods performance on inferring effective connectivity. ... The real resting-state f MRI dataset used in this paper are obtained from (Shah et al. 2017). ... We also use real task f MRI dataset by (Ramsey et al. 2010) to test the performance of the EC-GAN. |
| Dataset Splits | No | The paper mentions generating simulated data to select hyperparameters for EC-GAN, which implies a validation process: |
| Hardware Specification | No | No specific hardware details (e.g., GPU models, CPU models, cloud instance types, or memory) used for running the experiments are mentioned in the paper. |
| Software Dependencies | No | The paper describes the parameter configurations for comparison methods but does not provide specific software dependencies (e.g., library or solver names with version numbers) for the proposed EC-GAN or the experimental environment. |
| Experiment Setup | Yes | For a five nodes f MRI time series data, the hyper-parameters of EC-GAN are set as: the learning rate of generator and discriminator are 0.1, the number of units m is 100, sparsity parameter λ is 5. The threshold of causal parameters (A) is determined by the maximum number of parents Max P, and it is the same for all nodes. ... For the two simulated f MRI datasets (five nodes effective connectivity networks), we set Max P = 2. And for the two real f MRI datasets (eight and nine nodes effective connectivity networks), all hyper-parameters are the same with the simulated datasets but set Max P = 5. |