Brain-Computer Interface
Brain-Computer Interface (BCI) is a frontier field simultaneously unfolding across neuroscience, artificial intelligence, precision manufacturing, clinical medicine, and constitutional legislation. This chapter is not an introductory BCI textbook, but a systematic set of learning notes organized around the axis of "Intention-to-Action".
This chapter stands alongside Human-Like Intelligence. Human-Like Intelligence discusses "how to construct mind on the digital side" (predictive coding, world models, causality, meta-learning); this chapter discusses "how to establish a direct pathway between biological neurons and machines." Together they form two parallel paths toward AGI and embodied intelligence.
Chapter contents:
- Introduction — Why 2024-2026 is the tipping point for BCI × AI, reading paths
- 01 Foundations — BCI classification, history, core terminology and metrics
- 02 Neurophysiology — Motor cortex hierarchy, origins of neural signals, neural manifolds
- 03 Signal Acquisition — Invasive/minimally invasive/non-invasive electrodes, stimulation, preprocessing
- 04 Classical Decoding — Population vector, Kalman filter, ReFIT, linear discriminant
- 05 Deep Learning Decoders — EEGNet, LFADS, NDT, CEBRA, neural foundation models
- 06 Intention to Action ⭐ — I2A pipeline, shared autonomy, BCI→LLM→robot
- 07 Brain-to-Language — Invasive speech BCI, handwriting, non-invasive brain-to-text, LLM fusion
- 08 Brain-to-Image/Video — MindEye, MinD-Video, visual cortical prosthesis
- 09 Sensory Writing & Bidirectional BCI — ICMS, bidirectional BCI, memory prosthesis
- 10 Link to Embodied Intelligence ⭐ — Motor cortex as dynamical system, neural manifolds and RL policies
- 11 Commercial & Clinical Landscape — Neuralink, Synchron, Precision, Chinese BCI, regulation
- 12 Consumer & Non-Invasive — Muse/Emotiv, Apple neural sensors, thought typing
- 13 Ethics & Neurorights — Legislation, brain data privacy, AI alignment, UNESCO/EU
- 14 Datasets & Tools — NLB/FALCON, MNE/EEGLAB/CEBRA, learning resources