# Tau-progression forecasting on OASIS3 cross-cohort robustness (test)

Source: [Tau-progression forecasting on OASIS3 cross-cohort robustness (test)](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-16T08:49:51.290Z

Evaluation of models forecasting Tau progression using the OASIS3 cross-cohort robustness test set, assessed using metrics such as Mean Absolute Error (MAE), correlation metrics (ρ∆, Braak ρ), and Top-3 accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Tau-progression forecasting |
| Dataset (including any stated split) | OASIS3 cross-cohort robustness (test) |
| Metric | Mean Absolute Error (MAE) |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 15 of 15 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Mean Absolute Error (MAE)](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test.md?metric=Mean%20Absolute%20Error%20(MAE)>): Lower is better; unit: SCALAR.

- [Rho Delta (ρ∆)](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test.md?metric=Rho%20Delta%20(%CF%81%E2%88%86)>): Higher is better; unit: SCALAR.

- [Cosine Correlation (Cos.)](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test.md?metric=Cosine%20Correlation%20(Cos.)>): Higher is better; unit: SCALAR.

- [Top-3 Accuracy](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test.md?metric=Top-3%20Accuracy>): Higher is better; unit: PERCENT.

- [Braak Rho (Braak ρ)](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test.md?metric=Braak%20Rho%20(Braak%20%CF%81)>): Higher is better; unit: SCALAR.

- [Bias-corrected MAE (B-MAE)](<https://www.sota2.com/research/sota/tau-progression-forecasting-on-oasis3-cross-cohort-robustness-test.md?metric=Bias-corrected%20MAE%20(B-MAE)>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Mean Absolute Error (MAE) | Result date | Report paper |
| --- | --- | --- | --- |
| Ozdemir et al. | 1.35 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Giorgio et al. | 1.39 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Ren et al. | 1.59 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Karlsson et al. | 1.64 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Tekkesinoglu et al. | 1.81 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Rathore et al. | 1.85 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| BN-LTE + ATN-Z tau-free | 2.25 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| BN-LTE | 2.28 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| BN-LTE + PCA-Z | 2.3 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Raj et al. | 3.05 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Iturria-Medina et al. | 3.05 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Schafer et al. | 3.05 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Nguyen et al. | 5.92 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Jasodanand et al. | 6.66 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |
| Tabarestani et al. | 8.92 | Not recorded | 6a30d8a2d33dfb23e783ccb9 |

### Ozdemir et al.

Mean Absolute Error (MAE): 1.35. Result ID: 6a310e11fbc373592c2d739f.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Graph learning"}. This record does not assert independent reproduction.

### Giorgio et al.

Mean Absolute Error (MAE): 1.39. Result ID: 6a310e11fbc373592c2d7399.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Machine Learning"}. This record does not assert independent reproduction.

### Ren et al.

Mean Absolute Error (MAE): 1.59. Result ID: 6a310e11fbc373592c2d73a1.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Graph Learning"}. This record does not assert independent reproduction.

### Karlsson et al.

Mean Absolute Error (MAE): 1.64. Result ID: 6a310e11fbc373592c2d739a.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Machine Learning"}. This record does not assert independent reproduction.

### Tekkesinoglu et al.

Mean Absolute Error (MAE): 1.81. Result ID: 6a310e11fbc373592c2d73a0.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Graph learning"}. This record does not assert independent reproduction.

### Rathore et al.

Mean Absolute Error (MAE): 1.85. Result ID: 6a310e11fbc373592c2d739b.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Machine Learning"}. This record does not assert independent reproduction.

### BN-LTE + ATN-Z tau-free

Mean Absolute Error (MAE): 2.25. Result ID: 6a310e11fbc373592c2d73a4.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Ours"}. This record does not assert independent reproduction.

### BN-LTE

Mean Absolute Error (MAE): 2.28. Result ID: 6a310e11fbc373592c2d73a2.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Ours"}. This record does not assert independent reproduction.

### BN-LTE + PCA-Z

Mean Absolute Error (MAE): 2.3. Result ID: 6a310e11fbc373592c2d73a3.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Ours"}. This record does not assert independent reproduction.

### Raj et al.

Mean Absolute Error (MAE): 3.05. Result ID: 6a310e11fbc373592c2d7396.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Biophysical"}. This record does not assert independent reproduction.

### Iturria-Medina et al.

Mean Absolute Error (MAE): 3.05. Result ID: 6a310e11fbc373592c2d7397.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Biophysical"}. This record does not assert independent reproduction.

### Schafer et al.

Mean Absolute Error (MAE): 3.05. Result ID: 6a310e11fbc373592c2d7398.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Bayesian/Biophysical"}. This record does not assert independent reproduction.

### Nguyen et al.

Mean Absolute Error (MAE): 5.92. Result ID: 6a310e11fbc373592c2d739d.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Deep learning"}. This record does not assert independent reproduction.

### Jasodanand et al.

Mean Absolute Error (MAE): 6.66. Result ID: 6a310e11fbc373592c2d739e.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Deep learning"}. This record does not assert independent reproduction.

### Tabarestani et al.

Mean Absolute Error (MAE): 8.92. Result ID: 6a310e11fbc373592c2d739c.

[Reporting paper](<https://www.sota2.com/research/paper/6a30d8a2d33dfb23e783ccb9.md>)

Recorded evaluation configuration: {"Type":"Deep learning"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
