NC vs. AD Classification on ADNI
96.8AccuracyVision Transformer
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Vision TransformerStudy=[12], Model=Vision Transformer2026.02 | 96.8 | — | — | — | — | — | — | |
| GMUArchitecture=GMU, Modality=Multimodal2026.06 | 95.47 | — | 94.31 | 95.41 | — | — | 96.73 | |
| 3D CNNsArchitecture=3D CNNs, Modality=Unimodal (MRI)2026.06 | 95.39 | — | — | — | — | — | — | |
| 2D&3D CNNsArchitecture=2D&3D CNNs, Modality=Multimodal2026.06 | 95 | — | 93.33 | 93 | — | — | 96.66 | |
| camADArchitecture=camAD, Modality=Multimodal2026.06 | 94.5 | — | 95 | 94.9 | — | — | 94.9 | |
| 3D Multiscale CNNArchitecture=3D Multiscale CNN, Modality=Multimodal2026.06 | 94.11 | — | 93.33 | — | — | — | 94.27 | |
| Trans-ResNetStudy=[19], Model=Trans-ResNet2026.02 | 93.85 | — | — | — | — | — | — | |
| PT DCNArchitecture=PT DCN, Modality=Multimodal2026.06 | 92.7 | — | 91.7 | 96.4 | — | — | 93.5 | |
| BrainAnytimeModality=T1+Flair+PET2026.05 | 92.4 | — | — | 97.5 | 87.2 | — | — | |
| GMN4ADvalidation=5-fold cross validation, optimization=grid search2026.06 | 92.37 | — | 88.26 | — | 89.02 | 90.36 | 94.74 | |
| BrainAnytimeModality=T1+Flair2026.05 | 91.9 | — | — | 97 | 81.3 | — | — | |
| BrainAnytimeModality=T1+T2+Flair+PET2026.05 | 91.9 | — | — | 95.2 | 78.3 | — | — | |
| SAM-Brain3DModality=T1+Flair2026.05 | 91.4 | — | — | 95.3 | 81.9 | — | — | |
| UG-FedDA2025.12 | 90.54 | — | — | 91.68 | — | — | — | |
| Flex-MoEModality=T1+Flair+PET2026.05 | 90.4 | — | — | 95.3 | 84.1 | — | — | |
| FuseMoEModality=T1+Flair2026.05 | 90.1 | — | — | 94.3 | 78.3 | — | — | |
| MoE-retrieverModality=T1+Flair2026.05 | 90.1 | — | — | 93.9 | 77.6 | — | — | |
| MENetModality=T1+Flair2026.05 | 89.6 | — | — | 95.4 | 73.5 | — | — | |
| BrainAnytimeModality=Average2026.05 | 89.4 | — | — | 94.7 | 82.4 | — | — | |
| Flex-MoEModality=T1+T2+Flair+PET2026.05 | 89.3 | — | — | 93.6 | 70.7 | — | — | |
| MoE-retrieverModality=T1+T2+Flair+PET2026.05 | 89.1 | — | — | 90.6 | 69.9 | — | — | |
| FILvalidation=5-fold cross validation, optimization=grid search2026.06 | 88.78 | — | 83.98 | — | 84.61 | 85.29 | 91.57 | |
| Flex-MoEModality=T1+Flair2026.05 | 88.3 | — | — | 95.5 | 75.9 | — | — | |
| FedADA2025.12 | 87.84 | — | 73.47 | 91.22 | — | — | — | |
| MCADModality=T1+Flair2026.05 | 87.8 | — | — | 93.9 | 67.9 | — | — | |
| Stacked AutoEncodersArchitecture=Stacked AutoEncoders, Modality=Multimodal2026.06 | 87.76 | — | 88.57 | — | — | — | 87.22 | |
| GCNvalidation=5-fold cross validation, optimization=grid search2026.06 | 87.73 | — | 80.69 | — | 82.82 | 85.12 | 91.84 | |
| GATvalidation=5-fold cross validation, optimization=grid search2026.06 | 87.69 | — | 80.49 | — | 82.76 | 85.23 | 91.9 | |
| GINvalidation=5-fold cross validation, optimization=grid search2026.06 | 87.65 | — | 80.98 | — | 82.82 | 84.89 | 91.58 | |
| BrainAnytimeModality=T12026.05 | 86.5 | — | — | 93.4 | 88.4 | — | — | |
| FuseMoEModality=T1+T2+Flair+PET2026.05 | 86.3 | — | — | 87.9 | 56 | — | — | |
| MoE-retrieverModality=T1+Flair+PET2026.05 | 84.8 | — | — | 92.5 | 73.1 | — | — | |
| Multiscale DNNArchitecture=Multiscale DNN, Modality=Multimodal2026.06 | 84.6 | — | 80.2 | — | — | — | 91.8 | |
| BLvalidation=5-fold cross validation, optimization=grid search, cross-graph feature representation learning=disabled2026.06 | 84.29 | — | 81.4 | — | 78.98 | 77.05 | 85.95 | |
| 3D Resnet50Modality=T12026.05 | 84.2 | — | — | 93.1 | 87.2 | — | — | |
| BrainAnytimeModality=T1+T2+Flair2026.05 | 84.2 | — | — | 90.6 | 77 | — | — | |
| Flex-MoEModality=Average2026.05 | 84.2 | — | — | 91 | 74.8 | — | — | |
| MoE-retrieverModality=Average2026.05 | 83.1 | — | — | 87.2 | 74.7 | — | — | |
| FuseMoEModality=T1+Flair+PET2026.05 | 82.3 | — | — | 92 | 63.7 | — | — | |
| MoE-retrieverModality=T1+T2+Flair2026.05 | 82.1 | — | — | 86.9 | 72.9 | — | — | |
| FuseMoEModality=Average2026.05 | 82.1 | — | — | 87.5 | 67.8 | — | — | |
| SAM-Brain3DModality=T12026.05 | 81.1 | — | — | 91.7 | 84.2 | — | — | |
| BrainIACModality=T1+Flair2026.05 | 80.2 | — | — | 87.8 | 58.9 | — | — | |
| Flex-MoEModality=T1+T2+Flair2026.05 | 77.7 | — | — | 88.8 | 61.8 | — | — | |
| FuseMoEModality=T1+T2+Flair2026.05 | 76.9 | — | — | 83 | 60.5 | — | — | |
| BrainIACModality=T1+T2+Flair2026.05 | 76.6 | — | — | 82.1 | 62.8 | — | — | |
| SAM-Brain3DModality=T1+T2+Flair2026.05 | 76.6 | — | — | 85.8 | 64.2 | — | — | |
| BrainIACModality=T12026.05 | 76.1 | — | — | 82.5 | 79.1 | — | — | |
| Flex-MoEModality=T12026.05 | 75.2 | — | — | 81.7 | 81.3 | — | — | |
| FuseMoEModality=T12026.05 | 74.8 | — | — | 80.2 | 80.4 | — | — | |
| M3TModality=T12026.05 | 70.3 | — | — | 81.2 | 75.9 | — | — | |
| MoE-retrieverModality=T12026.05 | 69.4 | — | — | 72 | 79.9 | — | — | |
| BrainFoundInput Modality=Multimodal MRI (T1, T2, FLAIR), Pretraining Data=10,000 volumes across 12 datasets2026.06 | — | — | — | 81 | — | — | — | |
| Random ForestArchitecture=Random Forest, Modality=Unimodal (MRI)2026.06 | — | — | 88.6 | — | — | — | 92 | |
| SL-MAEInput Modality=Single contrast MRI, Model Variant=Ours (best)2026.06 | — | — | — | 94.5 | — | — | — |