[KongchangAI]
Model

ScaleBrain

A scale-consistent pretraining framework for electroencephalography (EEG) foundation models. It employs multi-scale EEG tokenization and scale-specific reconstruction to co-design inputs and targets, and uses axial spatio-temporal attention to fuse multi-scale representations.

Timeline (last 90 days)

Nov 11

The researchers propose ScaleBrain, a scale-consistent pretraining framework that explicitly coordinates multi-scale representations from both the model input and the optimization objective.

Unverified50%
Nov 11

ScaleBrain introduces a multi-scale EEG tokenization method that represents the same EEG signal through analysis windows of different time spans and shared temporal anchors, so that the model input naturally possesses scale structure.

Unverified50%
Nov 11

ScaleBrain adopts scale-specific reconstruction as the self-supervised objective to ensure that each temporal resolution receives sufficient supervisory signal.

Unverified50%
Nov 11

ScaleBrain designs an axial spatio-temporal attention mechanism that performs context modeling on the fused multi-scale representations along both the spatial and temporal dimensions, capturing inter-electrode spatial correlations and temporal dynamic dependencies respectively.

Unverified50%
Nov 11

ScaleBrain was extensively evaluated on six public BCI benchmark datasets and compared with 12 representative baseline methods.

Unverified50%
Nov 11

ScaleBrain achieved the highest mean Balanced Accuracy on five datasets and ranked third on the remaining one dataset.

Unverified50%
Nov 11

ScaleBrain contains only 5.0M parameters.

Unverified50%
Nov 11

Axial attention has mature applications in video understanding and high-dimensional image generation, and ScaleBrain transfers it to EEG multi-scale representation fusion.

Unverified50%

All Facts (8)

Source Articles