Research map / Graph learning / Knowledge and temporal graphs
Learning Dynamic Graph Representation of Brain Connectome with Spatio-Temporal Attention
Byung-Hoon Kim, Jong Chul Ye, Jae-Jin Kim
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add orthogonal regularization (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 87.46 -> 88.20 | Improved |
| add spatial attention READOUT (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 86.55 -> 87.46 | Improved |
| add temporal attention Transformer encoder (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 85.64 -> 86.55 | Improved |
| add timestamp encoding to node features (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 82.34 -> 85.64 | Improved |
| use Schaefer 400 atlas instead of AAL (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 85.36 -> 88.20 | Improved |
| use Schaefer 400 atlas instead of Harvard-oxford (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 82.07 -> 88.20 | Improved |
| raise edge threshold from 30% to 40% (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 88.20 -> 89.02 | Improved |
| change window length from 50 to 25 (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 88.20 -> 85.45 | Hurt |
| replace SERO with TopKPooling scoring (ablation) | STAGIN-SERO · HCP-Rest gender classification | accuracy (%): 88.20 -> 77.02 | Hurt |
| use dynamic graph STAGIN-SERO instead of static GIN | GNN FC classifier · HCP-Rest gender classification | accuracy (%): 81.34 -> 88.20 | Improved |
| use STAGIN-SERO instead of ST-GCN | dynamic FC classifier · HCP-Task decoding | accuracy (%): 98.92 -> 99.19 | Improved |
| use dynamic FC instead of static FC graphs | GNN FC classifier · HCP-Task decoding | accuracy (%): 93.87 -> 99.19 | Improved |
Closest papers by meaning
- $\textit{NeuroPath}$: A Neural Pathway Transformer for Joining the Dots of Human Connectomes (NeurIPS 2024 · similarity 0.88)
- BrainMAP: Learning Multiple Activation Pathways in Brain Networks (AAAI 2025 · similarity 0.88)
- Functional Connectomes of Neural Networks (AAAI 2025 · similarity 0.87)
- Beyond Pairwise Connections: Extracting High-Order Functional Brain Network Structures under Global Constraints (NeurIPS 2025 · similarity 0.87)
- Long-range Brain Graph Transformer (NeurIPS 2024 · similarity 0.87)
- Neurodegenerative Brain Network Classification via Adaptive Diffusion with Temporal Regularization (ICML 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
