What masking geometry works best for EEG foundation models?
Paper • 2609.33487 • Published • 3
eeg-fm-masking
58 EEG foundation-model encoders, MAE & JEPA × 29 mask geometries. Paper, code & interactive results: https://pierregtch.github.io/eeg-fm-masking
-
OpenEEG-Bench Arena
🧠7Comparing EEG models in an open and reproducible way
Note The downstream benchmark used in this paper
brain-bzh/reve-dataset
Preview • Updated • 1.9k • 7Note Pre-training dataset used in this paper
-
PierreGtch/eeg-fm-masking_mae_r9cm_L2
Feature Extraction • 12.7M • Updated • 143 -
PierreGtch/eeg-fm-masking_jepa_r9cm_L2
Feature Extraction • 12.7M • Updated • 591 -
PierreGtch/eeg-fm-masking_mae_rone_L1
Feature Extraction • 12.7M • Updated • 46 -
PierreGtch/eeg-fm-masking_mae_rone_L2
Feature Extraction • 12.7M • Updated • 33 -
PierreGtch/eeg-fm-masking_mae_rone_L4
Feature Extraction • 12.7M • Updated • 29 -
PierreGtch/eeg-fm-masking_mae_rone_L8
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_rone_L16
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_rone_L33
Feature Extraction • 12.7M • Updated • 32 -
PierreGtch/eeg-fm-masking_mae_r6cm_L1
Feature Extraction • 12.7M • Updated • 36 -
PierreGtch/eeg-fm-masking_mae_r6cm_L2
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_r6cm_L4
Feature Extraction • 12.7M • Updated • 35 -
PierreGtch/eeg-fm-masking_mae_r6cm_L8
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_mae_r6cm_L16
Feature Extraction • 12.7M • Updated • 36 -
PierreGtch/eeg-fm-masking_mae_r6cm_L33
Feature Extraction • 12.7M • Updated • 37 -
PierreGtch/eeg-fm-masking_mae_r9cm_L1
Feature Extraction • 12.7M • Updated • 39 -
PierreGtch/eeg-fm-masking_mae_r9cm_L4
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_mae_r9cm_L8
Feature Extraction • 12.7M • Updated • 31 -
PierreGtch/eeg-fm-masking_mae_r9cm_L16
Feature Extraction • 12.7M • Updated • 37 -
PierreGtch/eeg-fm-masking_mae_r9cm_L33
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_r12cm_L1
Feature Extraction • 12.7M • Updated • 36 -
PierreGtch/eeg-fm-masking_mae_r12cm_L2
Feature Extraction • 12.7M • Updated • 29 -
PierreGtch/eeg-fm-masking_mae_r12cm_L4
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_r12cm_L8
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_r12cm_L16
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_mae_r12cm_L33
Feature Extraction • 12.7M • Updated • 41 -
PierreGtch/eeg-fm-masking_mae_rall_L1
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_rall_L2
Feature Extraction • 12.7M • Updated • 39 -
PierreGtch/eeg-fm-masking_mae_rall_L4
Feature Extraction • 12.7M • Updated • 37 -
PierreGtch/eeg-fm-masking_mae_rall_L8
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_mae_rall_L16
Feature Extraction • 12.7M • Updated • 39 -
PierreGtch/eeg-fm-masking_jepa_rone_L1
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_jepa_rone_L2
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_jepa_rone_L4
Feature Extraction • 12.7M • Updated • 31 -
PierreGtch/eeg-fm-masking_jepa_rone_L8
Feature Extraction • 12.7M • Updated • 33 -
PierreGtch/eeg-fm-masking_jepa_rone_L16
Feature Extraction • 12.7M • Updated • 28 -
PierreGtch/eeg-fm-masking_jepa_rone_L33
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_jepa_r6cm_L1
Feature Extraction • 12.7M • Updated • 37 -
PierreGtch/eeg-fm-masking_jepa_r6cm_L2
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_jepa_r6cm_L4
Feature Extraction • 12.7M • Updated • 30 -
PierreGtch/eeg-fm-masking_jepa_r6cm_L8
Feature Extraction • 12.7M • Updated • 30 -
PierreGtch/eeg-fm-masking_jepa_r6cm_L16
Feature Extraction • 12.7M • Updated • 32 -
PierreGtch/eeg-fm-masking_jepa_r6cm_L33
Feature Extraction • 12.7M • Updated • 28 -
PierreGtch/eeg-fm-masking_jepa_r9cm_L1
Feature Extraction • 12.7M • Updated • 33 -
PierreGtch/eeg-fm-masking_jepa_r9cm_L4
Feature Extraction • 12.7M • Updated • 34 -
PierreGtch/eeg-fm-masking_jepa_r9cm_L8
Feature Extraction • 12.7M • Updated • 31 -
PierreGtch/eeg-fm-masking_jepa_r9cm_L16
Feature Extraction • 12.7M • Updated • 37 -
PierreGtch/eeg-fm-masking_jepa_r9cm_L33
Feature Extraction • 12.7M • Updated • 32 -
PierreGtch/eeg-fm-masking_jepa_r12cm_L1
Feature Extraction • 12.7M • Updated • 35 -
PierreGtch/eeg-fm-masking_jepa_r12cm_L2
Feature Extraction • 12.7M • Updated • 41 -
PierreGtch/eeg-fm-masking_jepa_r12cm_L4
Feature Extraction • 12.7M • Updated • 43 -
PierreGtch/eeg-fm-masking_jepa_r12cm_L8
Feature Extraction • 12.7M • Updated • 35 -
PierreGtch/eeg-fm-masking_jepa_r12cm_L16
Feature Extraction • 12.7M • Updated • 37 -
PierreGtch/eeg-fm-masking_jepa_r12cm_L33
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_jepa_rall_L1
Feature Extraction • 12.7M • Updated • 51 -
PierreGtch/eeg-fm-masking_jepa_rall_L2
Feature Extraction • 12.7M • Updated • 38 -
PierreGtch/eeg-fm-masking_jepa_rall_L4
Feature Extraction • 12.7M • Updated • 44 -
PierreGtch/eeg-fm-masking_jepa_rall_L8
Feature Extraction • 12.7M • Updated • 46 -
PierreGtch/eeg-fm-masking_jepa_rall_L16
Feature Extraction • 12.7M • Updated • 40