Back to work
BCI Preparation2025In progress
Neural Decoder Lab
A reproducible pipeline for offline EEG decoding — attention-based decoders, benchmarked latency budgets, honest baselines.
PyTorchMNEEEG
What it is
A slow-growing lab notebook for offline EEG decoding. The goal isn't a paper — it's to internalize the full path from raw EDF files to a decoder I actually understand.
Current status
- Reproducing a small BCI Competition IV-2a baseline with MNE + PyTorch
- Writing down every preprocessing choice as a decision log
- Comparing attention-based decoders against simple CSP + LDA baselines — honestly
Why
BCI reads exciting from far away and difficult from close up. I'd rather learn the difficult version first, on public data, before pretending to have opinions.