ABIDE I
Unknown
1,112 resting-state fMRI datasets from 539 individuals with autism spectrum disorder (ASD) and 573 typically developing controls across 17 international sites. Multi-site neuroimaging data for autism classification and biomarker discovery.
Benchmark Stats
SOTA History
accuracy
accuracy
Higher is better
| Rank | Model | Source | Score | Year | Paper |
|---|---|---|---|---|---|
| 1 | plymouth-dl-model 98% on 884 participant subset. Highlights visual cortex regions. | Editorial | 98 | 2025 | Source |
| 2 | mcbert Multi-modal CNN-BERT with leave-one-site-out cross-validation. Combines brain MRI and meta-features. | Editorial | 93.4 | 2025 | Source |
| 3 | ae-fcn Autoencoder + FCN combining fMRI and sMRI data (Rakic et al., 2020). | Editorial | 85 | 2025 | Source |
| 4 | multi-atlas-dnn Multi-atlas deep neural network with hinge loss. 79.13% on augmented data. | Editorial | 78.07 | 2025 | Source |
| 5 | asd-swnet Shared-weight feature extraction network. Precision: 76.15%, Recall: 80.65%. | Editorial | 76.52 | 2025 | Source |
| 6 | maacnn Multi-attention CNN. AUC: 0.79 on ABIDE-I. | Editorial | 75.12 | 2025 | Source |
| 7 | al-negat Adversarial learning-based node-edge graph attention network. 1,007 subjects across 17 sites. | Editorial | 74.7 | 2025 | Source |
| 8 | braingnn ROI-aware graph convolutional network. Interpretable biomarker discovery. 1,035 subjects. | Editorial | 73.3 | 2025 | Source |
| 9 | gcn Graph Convolutional Network combining fMRI and sMRI with max voting. | Editorial | 72.2 | 2025 | Source |
| 10 | multi-task-transformer Multi-task transformer neural network on UM dataset. Attention mechanism for feature extraction. | Editorial | 72 | 2025 | Source |
| 11 | phgcl-ddgformer Graph contrastive learning with graph transformer. 74.8% sensitivity. | Editorial | 70.9 | 2025 | Source |
| 12 | svm-connectivity Support Vector Machine with functional connectivity features. Classic baseline comparison. | Editorial | 70.1 | 2025 | Source |
| 13 | deep-learning-heinsfeld Deep learning approach by Heinsfeld et al. (2017). Anterior-posterior brain connectivity disruption. | Editorial | 70 | 2025 | Source |
| 14 | mvs-gcn Multi-view Site Graph Convolutional Network handling multi-site variability. | Editorial | 69.38 | 2025 | Source |
| 15 | abraham-connectomes Participant-specific functional connectivity matrices (Abraham et al., 2017). | Editorial | 67 | 2025 | Source |
| 16 | random-forest Random Forest baseline. Sensitivity: 69%, Specificity: 58%. | Editorial | 63 | 2025 | Source |
auc
auc
Higher is better
| Rank | Model | Source | Score | Year | Paper |
|---|---|---|---|---|---|
| 1 | asd-swnet AUC-ROC score for autism vs typical control classification. | Editorial | 81 | 2025 | Source |
| 2 | braingт Graph Transformer for brain disorder diagnosis. Significantly outperforms BrainNetTF (73.2%). | Editorial | 78.7 | 2025 | Source |
| 3 | gcn Best performing model in comprehensive comparison study (2024). | Editorial | 78 | 2025 | Source |
| 4 | svm-connectivity Traditional ML baseline with functional connectivity matrices. | Editorial | 77 | 2025 | Source |
| 5 | mvs-gcn Multi-view approach addressing site heterogeneity. | Editorial | 69.01 | 2025 | Source |