Introduction to Neurotechnology and Brain-Computer Interfaces (BCIs)
Key Takeaways:
-
Neurotechnology Overview: Engineering tech to monitor or modulate the nervous system. Interfaces can be invasive (single/multiple sites) or non-invasive (EEG, fMRI).
-
Neuron Basics: Brain has ~100B neurons; each with ~7000 synapses. Neural signals are binary (action potential or none) with EPSP/IPSP influencing postsynaptic neuron firing.
-
Recorded Signals & Applications:
- BCIs for communication and control
- Neurodegenerative disease monitoring (e.g., ALS)
- Visual prostheses
-
BCI Types & Examples:
- P300-based: Detects rare stimuli, used in spellers, wheelchairs, robotic arms
- SSVEP-based: Detects brain response to flickering stimuli; applications in spellers, AR integration
- Motor Imagery (MI): Uses C3/C4 electrodes for imagined movements, controlling robotic arms or wheelchairs
-
Generative AI Integration: GANs can reduce calibration time for P300- and MI-based BCIs
-
Industrial Applications: Nissan (brain-to-vehicle), Facebook CTRL Labs, Samsung TV control, Neuralink
-
Challenges: Early diagnosis of ALS, EEG accessibility for African hair types, user engagement, visual fatigue
Key Takeaways:
-
BCI System Components: Signal acquisition → preprocessing → feature extraction → classification → application → biofeedback
-
EEG Basics:
- Non-invasive; high temporal resolution, low spatial precision
- Signals amplified and filtered before visualization
- 10–20 system for electrode placement; supports 32–512 electrodes
-
Signal Acquisition & Electrode Types:
- Dry and wet electrodes, gel or pre-gelled, caps for hygiene and practicality
- Correct electrode-scalp contact is crucial for accurate signals
-
BCI Approaches:
- Active vs Natural Intent, Operant Conditioning vs Pattern Recognition
- Synchronous vs Asynchronous, Offline vs Online, Non-invasive vs Invasive
-
BCI Paradigms:
- Selective attention: P300, SSVEP
- Spontaneous signal: SMR, MI, SCP, mental tasks
-
EEG-based Signals:
- P300: Positive deflection ~300ms after stimulus
- SSVEP: Brain response at stimulus flicker frequency
- MI: ERD/ERS patterns during imagined/real movement
-
Experiment Design Notes:
- Controlled baseline vs task conditions
- Behavioral task engagement is key
- Pre-tests and pilot tests are essential
-
BCI Illiteracy: Causes include individual brain differences, system limitations, insufficient training, and psychological factors
Key Takeaways:
-
Hardware Used: g.tec Unicorn Hybrid Black, 8 dry electrodes (Fz, C3, Cz, C4, Pz, PO7, Oz, PO8), wireless EEG recording via Bluetooth
-
Signal Processing: Amplification, filtering, digitalization, extraction of SSVEP components
-
SSVEP Paradigm:
- Flickering stimuli at multiple frequencies on FHD screen (240 Hz refresh)
- Gray-scale flicker, green background, blue cue color
-
Software: Customizable scripts for band-pass filtering and SSVEP extraction, GitHub repo for reference
-
BCI Approaches Supported: P300, MI, SSVEP, code-based VEP
-
Practical Tips:
- Electrode placement and impedance checking critical
- Monitor participant attention and fatigue
- Ensure correct reference and ground electrode fixation
Summary of Notes for Day 1:
- Understand neurotechnology foundations (neurons, signals, interfaces, applications).
- Know EEG-based BCI paradigms (P300, SSVEP, MI) and how they relate to selective attention or spontaneous signals.
- Familiarize with hardware/software (Unicorn Hybrid Black, OpenBCI, electrode types, signal acquisition).
- Experiment design skills: Pre-tests, pilot tests, trial/run/epoch/event definitions, participant criteria.
- Practical lab skills: Electrode placement, EEG signal acquisition, handling SSVEP tasks, monitoring attention/fatigue.
- Applications awareness: Clinical (ALS, vision restoration), industrial (robotic arm, wheelchair, automotive), AI integration for calibration.