Repository
brain2qwerty
Non-invasive decoding of typed sentences from MEG and EEG brain recordings using a convolutional encoder, transformer, and character-level language model.
README
Brain2Qwerty
Decoding Sentences from Non-Invasive Recordings of the Brain.
Documentation
Publications
-
Non-invasive decoding of typed sentences from human brain activity (Nature Neuroscience, 2026).
-
Accurate Decoding of Natural Sentences from Non-Invasive Brain Recordings (preprint, 2026).
Code
Infrastructure
License
The code is released under CC BY-NC 4.0.
Data
- Brain2Qwerty v1: https://huggingface.co/datasets/bcbl190626/SpanishBCBL
- Brain2Qwerty v2: under embargo until the paper acceptation.
The datasets are collected by and belong to the BCBL — Basque Center on Cognition, Brain and Language.