Behavioral prediction on PNC (10-fold cross-val)
0.14Internalizing (Int.) PCCNERVE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| NERVEProtocol=SSL pre-training + downstream KRR, Input=FC matrices2026.05 | 0.14 | 0.13 | 0.12 | |
| GraphMAEProtocol=SSL pre-training + downstream KRR, Input=FC matrices2026.05 | 0.09 | 0.08 | 0.09 | |
| BrainGNNProtocol=Supervised end-to-end, Input=FC matrices2026.05 | 0.08 | 0.04 | 0.06 | |
| BrainNetTFProtocol=Supervised end-to-end, Input=BOLD time series2026.05 | 0.08 | 0.07 | 0.05 | |
| BrainNetCNNProtocol=Supervised end-to-end, Input=FC matrices2026.05 | 0.07 | 0.08 | 0.05 | |
| GATEProtocol=SSL pre-training + downstream KRR, Input=BOLD time series2026.05 | 0.07 | 0.08 | 0.09 | |
| BrainGSLsProtocol=SSL pre-training + downstream KRR, Input=FC matrices2026.05 | 0.05 | 0.08 | 0.04 | |
| BrainMassProtocol=SSL pre-training + downstream KRR, Input=FC matrices2026.05 | 0.03 | 0.09 | 0.1 |