Research map / Continual learning / Online and replay-based
Continual Learning for fMRI-Based Brain Disorder Diagnosis via Functional Connectivity Matrices Generative Replay
Qianyu Chen, Shujian Yu
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 graph knowledge distillation (GKD) (ablation) | FORGE · ASD | AAA: 0.683 -> 0.730 | Improved |
| add graph knowledge distillation (GKD) (ablation) | FORGE · BSNIP | AAA: 0.637 -> 0.685 | Improved |
| add logits knowledge distillation (LKD) (ablation) | FORGE · MDD | AAA: 0.695 -> 0.738 | Improved |
| add logits knowledge distillation (LKD) (ablation) | FORGE · ASD | FOR: 0.151 -> 0.116 | Improved |
| add hierarchical contextual Thompson sampling (HCTS) (ablation) | FORGE · ASD | AAA: 0.676 -> 0.730 | Improved |
| add HCTS (ablation) | FORGE · BSNIP | FOR: 0.137 -> 0.101 | Improved |
| add HCTS (ablation) | FORGE · MDD | AAA: 0.685 -> 0.738 | Improved |
| use FCM-VAE generator for augmentation instead of no augmentation | GCN · BSNIP | accuracy: 0.688 -> 0.727 | Improved |
| use FCM-VAE vs GR-SPD-GAN for augmentation | GCN · ASD | accuracy: 0.780 -> 0.783 | No clear effect |
| use ReGate generator for augmentation | GCN · ASD / BSNIP / MDD | accuracy: 0.762 -> 0.724; 0.688 -> 0.639; 0.756 -> 0.724 | Hurt |
| increase replay buffer from 256 to 384 (ablation) | FORGE · ASD | AAA: 0.730 -> 0.719 | Hurt |
| increase replay buffer from 128 to 256 (ablation) | FORGE · ASD | AAA: 0.685 -> 0.730 | Improved |
Closest papers by meaning
- Large Connectome Model: An fMRI Foundation Model of Brain Connectomes Empowered by Brain-Environment Interaction in Multitask Learning Landscape (AAAI 2026 · similarity 0.86)
- Task-Free Continual Generation and Representation Learning via Dynamic Expansionable Memory Cluster (AAAI 2024 · similarity 0.84)
- A Unified and General Framework for Continual Learning (ICLR 2024 · similarity 0.84)
- FR²Seg: Continual Segmentation Across Multiple Sites via Fourier Style Replay and Adaptive Consistency Regularization (AAAI 2025 · similarity 0.84)
- A scalable generative model for dynamical system reconstruction from neuroimaging data (NeurIPS 2024 · similarity 0.84)
- BrainWorld: A Structural-Prior-Conditioned Generative Model for Whole-Brain 4D fMRI Dynamics (NeurIPS 2026 · similarity 0.84)
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.
