===============================================================================
NEUROPHYSIOLOGY & SENSOR TECHNOLOGIES IN BCI — COMPREHENSIVE CURRICULUM
(As Covered Across MIT, Stanford, Harvard, UC Berkeley, Princeton,

Caltech, Yale, Imperial College London, ETH Zurich, CMU, Johns Hopkins)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 1: NEUROPHYSIOLOGY — HOW THE BRAIN GENERATES RECORDABLE SIGNALS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

1.1 CELLULAR FOUNDATIONS
─────────────────────────
• Neuron anatomy: soma, dendrites, axon, myelin sheath, synapses
• Resting membrane potential (~ −70 mV) — Na⁺/K⁺ pump, leak channels
• Action potential generation:
- Threshold (~ −55 mV)
- Depolarization (Na⁺ influx)
- Repolarization (K⁺ efflux)
- Hyperpolarization & refractory periods
• Synaptic transmission:
- Excitatory postsynaptic potentials (EPSPs)
- Inhibitory postsynaptic potentials (IPSPs)
- Temporal & spatial summation
• Glial cells & their role in signal modulation

1.2 NEURAL CODING PRINCIPLES
──────────────────────────────
• Rate coding: firing frequency encodes stimulus intensity
• Temporal coding: precise spike timing carries information
• Population coding: distributed ensembles encode complex variables
• Oscillatory coding: brain rhythms (delta, theta, alpha, beta, gamma)
• Neural plasticity & Hebbian learning ("fire together, wire together")

1.3 BRAIN RHYTHMS (EEG FREQUENCY BANDS)
─────────────────────────────────────────
┌──────────┬───────────┬──────────────────────────────────────────────┐
│ Band │ Range │ Primary Association │
├──────────┼───────────┼──────────────────────────────────────────────┤
│ Delta │ 0.5–4 Hz │ Deep sleep, unconsciousness │
│ Theta │ 4–8 Hz │ Memory, navigation, drowsiness │
│ Alpha │ 8–13 Hz │ Relaxation, eyes-closed, inhibition │
│ Beta │ 13–30 Hz │ Active concentration, motor planning │
│ Gamma │ 30–100 Hz │ Higher cognition, perception, binding │
└──────────┴───────────┴──────────────────────────────────────────────┘

1.4 FUNCTIONAL BRAIN REGIONS RELEVANT TO BCI
──────────────────────────────────────────────
• Motor cortex (M1) — voluntary movement, prosthetic control
• Premotor & supplementary motor areas (PMA/SMA) — movement planning
• Somatosensory cortex (S1) — tactile feedback, sensory BCI
• Broca's & Wernicke's areas — speech production & comprehension
• Occipital cortex (V1) — visual processing, P300/SSVEP
• Prefrontal cortex — decision-making, executive function
• Hippocampus — memory formation, spatial navigation
• Thalamus — relay station, deep brain stimulation target

1.5 NEUROVASCULAR COUPLING
──────────────────────────
• Relationship between neural activity & blood flow
• Hemodynamic response function (HRF): onset ~2s, peak ~5–6s
• Basis for fMRI BOLD signal & fNIRS measurements
• Limitation: indirect measure of neural activity, slow temporal dynamics

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 2: SENSOR TECHNOLOGIES — SIGNAL ACQUISITION METHODS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

2.1 NON-INVASIVE: EEG (ELECTROENCEPHALOGRAPHY)
───────────────────────────────────────────────
Physics:
• Measures scalp potential differences from summed post-synaptic
potentials of cortical pyramidal cells
• Signal attenuated & smeared by skull (~high resistivity), CSF, scalp

Electrode Types:
┌─────────────────┬─────────────────────────────────────────────────┐
│ Wet (Ag/AgCl) │ Gold standard; gel reduces impedance (<5 kΩ); │
│ │ time-consuming setup, uncomfortable for long use │
├─────────────────┼─────────────────────────────────────────────────┤
│ Dry │ No gel needed; pin/bristle design; higher │
│ │ impedance (>50 kΩ); faster setup, lower comfort │
├─────────────────┼─────────────────────────────────────────────────┤
│ Semi-dry │ Small saline reservoir; compromise between wet │
│ │ and dry; emerging consumer-grade option │
└─────────────────┴─────────────────────────────────────────────────┘

Electrode Placement:
• 10-20 International System (standard 21 electrodes)
• 10-10 Extended System (up to 71 electrodes)
• 10-5 High-Density System (up to 345 electrodes)
• Reference & ground electrode configuration (linked mastoids,
averaged reference, Cz reference)

Specifications:
• Spatial resolution: ~1–3 cm (poor, due to skull smearing)
• Temporal resolution: ~1 ms (excellent)
• Signal amplitude: 10–100 µV (very small; requires amplification)
• Sampling rate: typically 250–1000 Hz
• Channel count: 1–256+

Common EEG-BCI Paradigms:
• Motor imagery (MI): imagining left/right hand, feet movement
• P300 evoked potential: oddball paradigm for spellers
• Steady-state visual evoked potential (SSVEP): flickering stimuli
• Event-related desynchronization/synchronization (ERD/ERS)
• Slow cortical potentials (SCP)

2.2 NON-INVASIVE: fNIRS (FUNCTIONAL NEAR-INFRARED SPECTROSCOPY)
──────────────────────────────────────────────────────────────────
Physics:
• Near-infrared light (650–950 nm) penetrates scalp & skull
• Differential absorption by oxy-hemoglobin (HbO) & deoxy-hemoglobin
(HbR) — modified Beer-Lambert law
• Measures hemodynamic response (indirect neural activity proxy)

Hardware:
• Optodes: light emitters (LEDs or lasers) + detectors
• Pair spacing: typically 30 mm (measures ~15–25 mm depth)
• Configuration: single-distance, multi-distance, time-domain (TD),
frequency-domain (FD) systems

Specifications:
• Spatial resolution: ~1–3 cm (limited penetration depth ~1.5 cm)
• Temporal resolution: ~1–5 s (limited by hemodynamics)
• Portable, relatively motion-tolerant
• Often paired with EEG for hybrid BCI

Strengths:
• Less sensitive to motion artifacts than EEG
• Better spatial localization than EEG (for superficial cortex)
• Portable & lower cost than fMRI/MEG

2.3 NON-INVASIVE: MEG (MAGNETOENCEPHALOGRAPHY)
──────────────────────────────────────────────────
Physics:
• Detects magnetic fields (~10–1000 fT) generated by intra-cellular
currents in pyramidal neurons
• Magnetic fields pass undistorted through skull (unlike EEG)

Sensors:
┌──────────────────────────────────┬──────────────────────────────┐
│ SQUID (Superconducting Quantum │ Traditional; requires liquid │
│ Interference Device) │ helium cooling (−269°C); │
│ │ expensive, fixed installation │
├──────────────────────────────────┼──────────────────────────────┤
│ OPM (Optically Pumped │ Room-temperature operation; │
│ Magnetometer) │ wearable MEG systems; newer │
│ │ & more flexible │
└──────────────────────────────────┴──────────────────────────────┘

Specifications:
• Spatial resolution: ~2–5 mm (better than EEG, source-dependent)
• Temporal resolution: ~1 ms
• Field sensitivity: ~10 fT (extremely weak signals)
• Fixed installation (SQUID); wearable (OPM)

2.4 NON-INVASIVE: fMRI (FUNCTIONAL MAGNETIC RESONANCE IMAGING)
──────────────────────────────────────────────────────────────────
Physics:
• BOLD (Blood Oxygen Level Dependent) contrast
• Deoxy-hemoglobin is paramagnetic; oxy-hemoglobin is not
• Measures ratio changes (not absolute values)
• Requires strong static magnetic field (1.5T, 3T, or 7T)

Specifications:
• Spatial resolution: 1–3 mm (whole brain, excellent)
• Temporal resolution: 1–5 s (limited by hemodynamics)
• Not truly portable; extremely expensive
• Rarely used for real-time BCI (slow + immobile)

2.5 SEMI-INVASIVE: ECoG (ELECTROCORTICOGRAPHY)
──────────────────────────────────────────────────
Placement:
• Subdural: electrodes placed directly on exposed brain surface
(beneath dura mater) — requires craniotomy
• Epidural: electrodes placed between skull & dura mater
(no dural opening) — slightly safer, lower signal quality

Electrode Types:
• Macro-electrodes: disc/grid/strip electrodes (4 mm diameter,
10 mm spacing), stainless steel or platinum
• Micro-ECoG: higher-density arrays (1–2 mm pitch)
• High-density ECoG grids (up to 256+ contacts)

Signal Characteristics:
• Records local field potentials (LFPs) — summed synaptic activity
• Does NOT record single-unit spikes (too far from neurons)
• Higher amplitude than scalp EEG (~100–1000 µV)
• Broader bandwidth (DC to ~500 Hz; can capture high gamma)
• Better spatial resolution than EEG (~1–5 mm)
• Less susceptible to artifacts (eye blinks, muscle)

Clinical Context:
• Typically used in epilepsy patients undergoing pre-surgical
monitoring (electrodes already implanted for clinical reasons)
• Golden opportunity for BCI research in human subjects

2.6 INVASIVE: INTRACORTICAL MICROELECTRODE ARRAYS
─────────────────────────────────────────────────────

┌─────────────────────────────────────────────────────────────────┐
│ a) UTAH ARRAY (Blackrock Neurotech) │
├─────────────────────────────────────────────────────────────────┤
│ Structure: │
│ • 10×10 grid of silicon micro-needles │
│ • Needle length: 1–1.5 mm │
│ • Electrode spacing: 400 µm │
│ • Up to 96 active recording channels per array │
│ • Expandable to 1024 channels in multi-array systems │
│ │
│ Electrode Coatings: │
│ • Platinum (Pt) — recording only │
│ • Sputtered iridium oxide (IrOx) — recording + stimulation │
│ │
│ Key Specs: │
│ • Records single-unit & multi-unit activity (spikes) │
│ • Signal-to-noise ratio: ~5–10 dB │
│ • FDA-cleared for chronic human use (NeuroPort) │
│ • Documented stability: up to 8+ years in human implants │
│ • Gold standard for human intracortical BCI research │
│ │
│ Limitations: │
│ • Foreign body response → glial scarring over time │
│ • Tissue encapsulation reduces signal quality │
│ • Material degradation with long implantation │
│ • Rigid silicon substrate can cause mechanical damage │
│ │
│ Source: blackrockneurotech.com/products/utah-array │
└─────────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────┐
│ b) NEUROPIXELS PROBES │
├─────────────────────────────────────────────────────────────────┤
│ Structure: │
│ • Single-shank silicon probe │
│ • Shank length: 10 mm, width: 70 µm │
│ • 960 recording sites along shaft │
│ • 20 µm site-to-site spacing │
│ • 384 simultaneously recordable channels │
│ • Sampling rate: ≥30 kHz per channel │
│ │
│ Electrode Coating: │
│ • Titanium nitride (TiN) │
│ • Low impedance (~1 MΩ) │
│ • Very high SNR (>10 dB) │
│ │
│ Key Advantages: │
│ • Ultra-high channel density (currently highest available) │
│ • Flexible recording configuration (select any 384 of 960) │
│ • On-probe amplification & multiplexing (reduces wire count) │
│ • Can record across multiple brain regions in one insertion │
│ │
│ Status: │
│ • Primarily used in animal research │
│ • Emerging translational & potential human trials │
│ • Not yet FDA-cleared for chronic human use │
└─────────────────────────────────────────────────────────────────┘

┌─────────────────────────────────────────────────────────────────┐
│ c) EMERGING INTRACORTICAL DESIGNS │
├─────────────────────────────────────────────────────────────────┤
│ │
│ Neuralink (Neuralink Corp.): │
│ • Thread-based flexible polymer electrodes │
│ • Inserted via surgical robot (avoids blood vessels) │
│ • 1024 electrodes per thread bundle (early versions) │
│ • Wireless data transmission (Bluetooth Low Energy) │
│ • FDA approval for first-in-human trial (2023) │
│ │
│ Michigan-Style Probes: │
│ • Planar shanks with multiple recording sites along the edge │
│ • Can span multiple cortical layers │
│ • Can combine recording + stimulation sites │
│ • Flexible polymer variants reduce tissue damage │
│ │
│ Neural Lace / Mesh Electronics: │
│ • Ultra-flexible mesh that integrates with brain tissue │
│ • Syringe-injectable │
│ • Minimal foreign body response (most biocompatible) │
│ • Very early stage; limited channel counts │
│ │
│ Stentrode (Synchron): │
│ • Deployed via vasculature (jugular vein) │
│ • Sits near motor cortex inside a blood vessel │
│ • No craniotomy needed │
│ • 16 electrodes; records local field potentials │
│ • FDA Investigational Device Exemption granted │
│ • First-in-human US trials underway │
└─────────────────────────────────────────────────────────────────┘

2.7 PERIPHERAL & SPINAL INTERFACES
────────────────────────────────────
• Peripheral nerve electrodes (cuff, intrafascicular, regenerative)
• EMG-based muscle signal acquisition (surface & intramuscular)
• Spinal cord stimulation & recording epidural leads
• Vagus nerve stimulation (VNS) electrodes
• Used in: prosthetics, bladder/bowel control, pain management

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 3: COMPARATIVE SPECIFICATIONS — ALL MODALITIES
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

┌───────────────┬──────────────┬──────────────┬────────────┬───────────┬──────────┐
│ Modality │ Spatial Res. │ Temporal Res.│ Invasiveness│ Portability│ Cost │
├───────────────┼──────────────┼──────────────┼────────────┼───────────┼──────────┤
│ fMRI │ 1–3 mm │ 1–5 sec │ None │ Fixed │ $$$$$ │
│ MEG (SQUID) │ 2–5 mm │ ~1 ms │ None │ Fixed │ $$$$ │
│ MEG (OPM) │ 2–5 mm │ ~1 ms │ None │ Wearable │ $$$ │
│ fNIRS │ 1–3 cm │ 1–5 sec │ None │ Portable │ $$ │
│ EEG (scalp) │ 1–3 cm │ ~1 ms │ None │ Portable │ $ │
│ ECoG │ 1–5 mm │ ~1 ms │ Semi-invasive│ Surgical │ $$$ │
│ LFP (depth) │ ~1 mm │ ~1 ms │ Invasive │ Surgical │ $$$ │
│ Utah Array │ <100 µm │ ~1 ms │ Invasive │ Surgical │ $$$ │
│ Neuropixels │ <50 µm │ ~0.03 ms │ Invasive │ Surgical │ $$$ │
│ Neuralink │ ~50 µm │ ~0.03 ms │ Invasive │ Implantable│ $$$ │
│ Stentrode │ ~5 mm │ ~1 ms │ Min-invasive │ Implantable│ $$$ │
└───────────────┴──────────────┴──────────────┴────────────┴───────────┴──────────┘

Signal Amplitude Comparison:
Scalp EEG: ~10–100 µV
ECoG: ~100–1000 µV
Intracortical: ~50–500 µV (LFP); ~100 µV–10 mV (spikes)
MEG: ~10–1000 fT (femtotesla)

Frequency Bandwidth:
Scalp EEG: ~0.5–100 Hz
ECoG: ~DC–500 Hz (captures high gamma: 80–150 Hz)
Intracortical: DC–10 kHz (spike detection requires ≥30 kHz sampling)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 4: SIGNAL CONDITIONING & FRONT-END ELECTRONICS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

4.1 AMPLIFICATION CHAIN
──────────────────────────
Stage 1: Electrode-skin interface (impedance reduction via gel/preparation)
Stage 2: Preamplifier (high input impedance >100 MΩ, gain ~10–100×)
Stage 3: Instrumentation amplifier (differential, CMRR >100 dB)
Stage 4: Anti-aliasing filter (low-pass before ADC)
Stage 5: Analog-to-digital converter (ADC, typically 16–24 bit)

4.2 NOISE & ARTIFACT SOURCES
──────────────────────────────
┌────────────────────┬───────────────────────────────────────────────┐
│ Source │ Characteristics │
├────────────────────┼───────────────────────────────────────────────┤
│ Power line (50/60Hz)│ Narrowband sinusoidal; removable with notch │
│ │ or comb filter │
├────────────────────┼───────────────────────────────────────────────┤
│ Electrode pop │ Sudden baseline shift from gel movement/contact │
├────────────────────┼───────────────────────────────────────────────┤
│ Eye blink (EOG) │ Large low-frequency transient (~100–200 µV) │
├────────────────────┼───────────────────────────────────────────────┤
│ Muscle (EMG) │ High-frequency broadband (>30 Hz, 50–500 µV) │
├────────────────────┼───────────────────────────────────────────────┤
│ Cardiac (ECG) │ Periodic QRS complex leakage (~1 mV) │
├────────────────────┼───────────────────────────────────────────────┤
│ Motion artifact │ Baseline drift from cable/electrode movement │
├────────────────────┼───────────────────────────────────────────────┤
│ Thermal noise │ Johnson-Nyquist from electrode impedance │
└────────────────────┴───────────────────────────────────────────────┘

4.3 COMMON FILTERING STRATEGIES
─────────────────────────────────
• Bandpass filter: typically 0.5–40 Hz (or 0.5–100 Hz for ECoG)
• Notch filter: 50 Hz (Europe) or 60 Hz (USA) for powerline noise
• Common Average Referencing (CAR) & Laplacian spatial filtering
• Independent Component Analysis (ICA) for artifact removal
• Adaptive filtering for real-time artifact cancellation

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 5: HYBRID & MULTIMODAL SYSTEMS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

5.1 EEG + fNIRS HYBRID BCI
───────────────────────────
• Combines fast temporal resolution (EEG) with better spatial
localization (fNIRS)
• Common features: EEG ERD/ERS + fNIRS HbO slope indicators
• Shared spatial filtering (CSP applied to both modalities)
• Applications: motor imagery classification, quadcopter/drone control,
neurorehabilitation
• Challenge: temporal alignment — EEG is ms-level, fNIRS has ~5s delay

5.2 EEG + EOG / EMG HYBRID
────────────────────────────
• Eye movements (EOG) provide intentional control channels
• Muscle signals (EMG) can augment or replace degraded brain signals
• Useful for users with residual motor function
• Approach: weighted fusion or hierarchical classification

5.3 INVASIVE RECORDING + ELECTRICAL STIMULATION
─────────────────────────────────────────────────
• Bidirectional BCI: recording + microstimulation
• Somatosensory feedback via intracortical microstimulation (ICMS)
in S1
• Closed-loop embodiment for prosthetic limbs
• Utah array supports stimulation with IrOx-coated tips

5.4 DATA SYNCHRONIZATION
─────────────────────────
• Lab Streaming Layer (LSL) — time-synced multi-stream acquisition
• Hardware triggers for cross-modality alignment
• Timestamp jitter requirements: <1 ms for closed-loop systems

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 6: SAFETY, BIOCOMPATIBILITY & REGULATORY CONSIDERATIONS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

6.1 NON-INVASIVE SAFETY
─────────────────────────
• EEG: minimal risk — skin irritation from gel, allergic reactions
• fNIRS: minimal risk — safe optical exposure levels
• MEG: minimal risk — passive measurement, no energy deposition

6.2 INVASIVE SAFETY & BIOCOMPATIBILITY
───────────────────────────────────────
• Foreign body response (FBR): glial scar formation, astrocyte
activation, neurodegeneration near electrode
• Strategies to mitigate FBR:
- Flexible polymer substrates (match brain mechanical properties)
- Drug-eluting coatings (dexamethasone, anti-inflammatory agents)
- Bioactive coatings (PEDOT, graphene, hydrogels)
- Ultra-small features (<10 µm) to minimize tissue displacement
• Sterilization: autoclave, ethylene oxide, gamma irradiation
• Long-term stability challenges: electrode delamination, tip
corrosion, insulator degradation

6.3 REGULATORY PATHWAYS
────────────────────────
┌──────────────┬──────────────────────────────────────────────────┐
│ FDA (USA) │ Investigational Device Exemption (IDE) → │
│ │ Premarket Approval (PMA) or 510(k) clearance │
├──────────────┼──────────────────────────────────────────────────┤
│ CE (EU) │ Conformité Européenne marking; requires clinical │
│ │ evaluation & ISO 13485 quality system │
├──────────────┼──────────────────────────────────────────────────┤
│ MHRA (UK) │ Post-Brexit UKCA marking; aligned with EU MDR │
├──────────────┼──────────────────────────────────────────────────┤
│ IMDRF │ International Medical Device Regulators Forum — │
│ │ harmonizing neurodevice classification globally │
└──────────────┴──────────────────────────────────────────────────┘

6.4 ETHICAL CONSIDERATIONS
───────────────────────────
• Neuroprivacy: who owns brain data?
• Cognitive liberty: right to mental self-determination
• Identity & agency: does BCI-mediated action feel like "your own"?
• Equity & access: will BCIs be affordable only to wealthy?
• Dual-use: military applications (cognitive enhancement, weaponization)
• Informed consent in severely paralyzed patients (communication barriers)
• Brain data security: vulnerability to hacking/neural malware

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 7: WHICH UNIVERSITIES TEACH WHAT
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

┌──────────────────────────┬──────────────────────────────────────────────┐
│ University │ Neurophysiology & Sensor Emphasis │
├──────────────────────────┼──────────────────────────────────────────────┤
│ MIT │ Neural signal biophysics; EEG/ECoG/MEG/ │
│ │ fNIRS; closed-loop design; Neuralink/Media │
│ │ Lab neural interfacing research │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Stanford │ Invasive micro-electrode arrays; speech BCI; │
│ │ motor decoding; clinical neuroprosthetics │
│ │ (Henderson, Willett labs) │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Harvard │ Soft/flexible bioelectronics; nanotechnology; │
│ │ optogenetics; brain stimulation; peripheral │
│ │ nerve interfaces │
├──────────────────────────┼──────────────────────────────────────────────┤
│ UC Berkeley │ EEG/EMG/ECoG hardware design; amplifiers; │
│ │ wireless acquisition; BCI2000/OpenBCI tools │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Princeton │ Neural coding; motor cortex physiology; │
│ │ recording & stimulation methods │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Caltech │ Electrophysiology; single-unit recording; │
│ │ optogenetics; systems neuroscience │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Yale │ Sensory & motor processing; neural systems; │
│ │ neurotechnology applications │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Imperial College London │ EEG/ECoG/implantable acquisition; artifact │
│ │ removal; real-time processing pipelines │
├──────────────────────────┼──────────────────────────────────────────────┤
│ ETH Zurich │ Neural information processing; optogenetics; │
│ │ wireless implantable devices; biocompatible │
│ │ electrodes │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Carnegie Mellon (CMU) │ EEG/ECoG/intracortical; electrode array │
│ │ design; BCI stabilization across sessions; │
│ │ neural modulation │
├──────────────────────────┼──────────────────────────────────────────────┤
│ Johns Hopkins │ EEG/ECoG/implantable recording & stimulation; │
│ │ clinical neurotechnology; precision surgery; │
│ │ neural implants (EN.580.742) │
└──────────────────────────┴──────────────────────────────────────────────┘

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 8: RECOMMENDED LAB COMPONENTS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Lab 1: EEG Acquisition Basics
→ Set up 10-20 cap, acquire resting-state EEG, identify frequency bands

Lab 2: Motor Imagery BCI
→ Record imagined left/right hand movement; extract ERD/ERS features;
train LDA classifier; visualize real-time cursor control

Lab 3: P300 Speller
→ Implement oddball paradigm; collect P300 epochs; classify target
vs. non-target; build matrix speller interface

Lab 4: SSVEP BCI
→ Design flickering LED stimulus; record SSVEP at target frequencies;
apply canonical correlation analysis (CCA) for classification

Lab 5: Signal Processing Pipeline
→ Load raw EEG; apply bandpass/notch filtering; ICA artifact removal;
epoching; feature extraction (PSD, CSP)

Lab 6: ECoG Signal Analysis (provided dataset)
→ Analyze human ECoG recordings; extract high-gamma activation;
compare with scalp EEG for same task

Lab 7: Intracortical Spike Sorting (provided dataset)
→ Load Utah array data; perform spike detection & sorting (PCA +
clustering); compute tuning curves for motor task

Lab 8: Hybrid EEG-fNIRS BCI
→ Simultaneously acquire EEG + fNIRS during motor imagery; fuse
features; compare single-modality vs. hybrid classification accuracy

Lab 9: Real-Time Closed-Loop BCI
→ Build end-to-end pipeline: acquisition → preprocessing → feature
extraction → classification → feedback display; target latency <500 ms

Lab 10: Capstone Design Project
→ Teams design, implement, and evaluate a complete BCI system for a
chosen application (communication, prosthetic control, gaming,
neurofeedback, or rehabilitation)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
PART 9: KEY REFERENCE RESOURCES
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

TEXTBOOKS:
• "Brain-Computer Interfaces: Principles and Practice" — Wolpaw & Wolpaw
• "Principles of Neural Science" (6th Ed.) — Kandel, Schwartz, Jessell
• "Neural Engineering" — He & Li
• "Bioelectricity: A Quantitative Approach" — Plonsey & Barr
• "Introduction to Neural Engineering for Motor Rehabilitation" —

DATASETS:
• PhysioNet (physionet.org) — EEG & physiological datasets
• OpenNeuro (openneuro.org) — Shared brain imaging data
• BNCI Horizon 2020 (bnci-horizon-2020.eu) — Benchmark BCI datasets
• UCSD EEG Lab — Open EEG processing tutorials
• CRCNS (crcns.org) — Collaborative research in computational neuroscience

OPEN-SOURCE SOFTWARE:
• OpenBCI — Hardware + GUI for DIY EEG/fNIRS acquisition
• BCI2000 — General-purpose research BCI platform
• BCILAB — MATLAB/Octave toolbox for BCI development
• MNE-Python — MEG/EEG analysis suite (mne.tools)
• EEGLAB — MATLAB toolbox for EEG processing (sccn.ucsd.edu/eeglab)
• Lab Streaming Layer (LSL) — Time-synced multi-stream data acquisition
• OpenViBE — Real-time BCI software platform

HARDWARE VENDORS:
• Blackrock Neurotech — Utah Array, Neural Signal Processor
• g.tec — EEG amplifiers, BCI research systems
• OpenBCI — Consumer-grade EEG/fNIRS hardware
• Paradromics — Connexus BCI platform
• Synchron — Stentrode vascular BCI
• Neuralink — Thread-based wireless BCI
• NeuroPaces — Responsive neurostimulation (epilepsy)
• Intan Technologies — Electrophysiology amplifiers & recording

PROFESSIONAL SOCIETIES:
• BCI Society (bcisociety.org) — International BCI meetings
• IEEE Engineering in Medicine and Biology Society (EMBS)
• Society for Neuroscience (SfN)
• Organization for Human Brain Mapping (OHBM)

===============================================================================

END OF NEUROPHYSIOLOGY & SENSOR TECHNOLOGIES