tribe-v2-bci-appliedlisted

# TRIBE v2 Applied BCI Skill Agentic skill for applied BCI research and neuro-informed content optimization — from predicting fMRI cortical responses to media without brain scanners, through stimulus optimization and attention ranking, to generating cortical priors for non-invasive BCI decoding research. > **License constraint**: TRIBE v2 is CC BY-NC 4.0. This skill is for non-commercial research only. Commercial neuromarketing, advertising optimization, or audience profiling for profit requires a separate license from Meta. Read [references/ethics-privacy.md](references/ethics-privacy.md) before any applied use. --- ## Quick Start ### 1. Install TRIBE v2 ```bash # Python 3.11+ required git clone https://github.com/facebookresearch/tribev2 cd tribev2 pip install -e . ``` ### 2. Load Model and Run First Prediction ```python from tribev2 import TribeModel # Load model — downloads weights on first run (~several GB) model = TribeModel.from_pretrained("facebook/tribev2", cache_folder="./cache") # Build events dataframe from your stimulus df = model.get_events_dataframe(video_path="path/to/video.mp4") # Predict cortical responses preds, segments = model.predict(events=df) # preds.shape = (n_timesteps, n_vertices) # n_vertices ~20,000 on fsaverage5 surface mesh print(f"Predicted response shape: {preds.shape}") print(f"Mean activation across all cortex: {preds.mean():.4f}") ``` ### 3. Supported Input Modalities ```python # Video (extracts visual + auditory + motion feat