--- library_name: samba-eeg pipeline_tag: feature-extraction tags: - eeg - brain-computer-interface - self-supervised-learning - mamba - pytorch --- # SAMBA EEG representation models. [Project website](https://samba.dcode.ai/) · [Code](https://github.com/Jiazhen-Hong/SAMBA) · [Paper](https://arxiv.org/abs/2511.18571) SAMBA was developed during my 2025 internship at Emotiv. The original checkpoints are no longer available, so we retrained both models. SAMBA-100s follows the original training method with a rebuilt TUAB pipeline and validation split. SAMBA-2s adds HBN, Muse, and DICE to the Emotiv data for greater electrode-layout diversity. ## Checkpoints Both models use 128 Hz EEG. The names indicate the pretraining window; the encoders support other input lengths and channel layouts. | | [SAMBA-2s](SAMBA-2s) | [SAMBA-100s](SAMBA-100s) | |---|---|---| | Window | 2 s, 256 samples | 100 s, 12,800 samples | | Data | Emotiv¹, HBN, Muse, DICE | TUAB | | Pretraining channels | 4, 14, 19, 129 | 16 bipolar | | Parameters | 1.06M | 1.06M | ¹ Alpha, Attention, Crowdsource, DriverDistraction, and STEW. ## Learned EEG spatial weights SAMBA uses electrode coordinates to learn how to combine EEG channels. These maps show how the spatial weights change during training. | Eyes open vs. closed | Driver distraction | Motor imagery | |:---:|:---:|:---:| | Eyes open vs. closed | Driver distraction | Motor imagery | ## Usage Requires a compatible CUDA system. ```bash git clone https://github.com/Jiazhen-Hong/SAMBA.git cd SAMBA pip install -e '.[mamba,hub]' ``` ```python from pathlib import Path import torch from huggingface_hub import snapshot_download from samba_eeg import SambaModel variant = "SAMBA-2s" # or "SAMBA-100s" folder = snapshot_download( "Jay1208/SAMBA", allow_patterns=[f"{variant}/*"], ) model = SambaModel.from_pretrained(Path(folder) / variant).cuda().eval() ``` Prepare EEG at 128 Hz: - `signals`: `[batch, channels, samples]`. - `positions`: `[channels, 3]` or `[batch, channels, 3]`, matching the channel order and the coordinate units and frame in `config.json`. - Normalize each channel in each window by its absolute 95th percentile. TUAB pretraining uses a 0.5–45 Hz bandpass. ```python signals = signals.to(device="cuda", dtype=torch.float32) positions = positions.to(device="cuda", dtype=torch.float32) # Skip if already normalized. scale = torch.quantile(signals.abs(), 0.95, dim=-1, keepdim=True) signals = signals / scale.clamp_min(1e-8) with torch.inference_mode(): features = model.encode(signals, positions) pooled_features = features.mean(dim=1) ``` `features` has shape `[batch, floor(samples / 8), 128]`. Use the pooled features for linear probing, or fine-tune the encoder. ## Citation ```bibtex @article{hong2026samba, title={SAMBA: Toward an EEG Foundation Model via Differential Mamba and Coordinate-Driven Spatial Embedding}, author={Hong, Jiazhen and Mackellar, Geoffrey and Ghane, Soheila}, journal={arXiv preprint arXiv:2511.18571}, year={2026} } ```