Technological evolution of BCI 1970s: research algorithms to reconstruct movements from motor cortex neurons 1980s: Johns Hopkins researchers found a mathematical relationship between electrical responses of single motor-cortex neurons in rhesus macaque monkeys and the direction that monkeys moved their arms (based on a cosine function). 1990s: Several groups able to capture complex brain motor centre signals using recordings from neurons and use these to control external devices The common thread throughout the research is the remarkable cortical plasticity of the brain, which often adapts to BCIs

Practical Elements Signal acquisition: the BCI system's recorded + digitised brain signal input. Signal processing: conversion of raw information into device command Feature extraction = the determination of a meaningful change in signal Feature translation = the conversion of that signal alteration to a device command. Statistical analysis on the basis of the probability function that an electrophysiological event correlates with a given cognitive or motor task. Device output: the overt command or control functions produced. word processing, communication, wheel chair, prosthetic limb. new output channel, therefore must have feedback to improve how they alter their electro physiological signal. Operating protocol: the manner in which the system is turned on/off.

Neurosurgical issues Safety Durability/Reliability: scar, removal and re-implantation Consistency, Useful complexity (DOF) Suitability Speed and accuracy Efficacy: Technical vs. Practical

EEG-based BCI P300 evoked potentials Sensorimotor Cortex Rhythms brain's response to infrequent or significant stimuli from its response to routine stimuli. Sensorimotor Cortex Rhythms Movement or preparation for movement is typically accompanied by a decrease in µ (8–12 Hz) and beta (18–26 Hz) activity over sensorimotor cortex People, including those with ALS or SCI have learned to control µ or beta amplitudes in the absence of movement or sensation susceptible to external forces (i.e., electrode movement) and contamination less fidelity and spatial specificity and a limited frequency detection (<40Hz), resulting in prolonged user training for higher levels of control. Spatial and frequency limitations prohibits complexity of movements supported by EEG

Single unit-based BCI Best signal for BCI control has been achieved with multiple, single-unit action potentials recorded in parallel directly from cerebral cortex, in terms of accuracy, speed and DOF than single unit data. Obtaining long-term stability of single unit recordings has proven difficult - Only provide a year of BCI control Require insertion of a recording electrode into the brain parenchyma Prone to scarring, implanted in eloquent regions of cortex

Simultaneously recorded intracellular and extracellular signals

Spatial dependence of spike waveform

Chronic recording electronics

How do you know you have a single unit? Spike train autocorrelation Always verify refractory period relative to long-time asymptote of autocorrelation function Latency Variability Analysis

Antidromic Spike Collision

Multiple neural signals Time (sec) Voltage (A/D Levels) Time (sec) Voltage (A/D Levels) 3 msec

Spike sorting Raw Data Spike Detector Region from previous slide Neuron #1 Spikes Neuron #2 Spikes Time (sec)

The ‘graduate student’ algorithm Time (sec) Voltage (A/D Levels) Raw Data Threshold detector at 32 Time (msec) Voltage (A/D Levels) Candidate Waveforms # of Intervals Time (msec) Interspike Interval Histogram Spike Height vs. Width Plot Width (msec) Height (A/D Levels)

General framework Locate Spikes Preprocess Waveforms Density Estimation Spike Classification Quality Measures

Principal Component Analysis Create “feature vector” for each spike “Feature space”

Post-Stimulus Time Histogram (PSTH) 40 µV 200 ms

Semi-automatic Clustering

Electrocorticogram-based system Electrocorticogram (EcoG) measures electrical activity of the brain taken from beneath the skull (subdural or epidural) Gamma rhythms as well as µ and beta rhythms are prominent in ECoG during movements higher spatial resolution, better signal-to-noise ratio, wider frequency range, and lesser training requirements than scalp-recorded EEG

Electric current dipole fields

Power spectrum handling & auto correlation function

Artifact detection in EEG EOG EMG

Stimulation mapping to locate cortical areas

Electromyography Indicator for muscle activation/deactivation Electrode Categories Inserted Fine-wire (Intra-muscular) Needle Surface

Fine wire electrode vs. surface electrode Pros Extremely sensitive Record single muscle activity Access to deep musculature Little cross-talk concern Cons Requires medical personnel, certification Repositioning nearly impossible Detection area may not be representative of entire muscle Pros Quick, easy to apply No medical supervision, required certification Minimal discomfort Cons Generally used only for superficial muscles Cross-talk concerns No standard electrode placement May affect movement patterns of subject Limitations with recording dynamic muscle activity

Characteristics of EMG signal and noise Amplitude range: 0–10 mV (+5 to -5) prior to amplification Usable energy: Range of 0 - 500 Hz Dominant energy: 50 – 150 Hz Noise Inherent noise in electronics equipment: 0 – thousands Hz Cannot be eliminated Reduced by using high quality components Ambient noise: e.g. 60Hz Radio transmission, electrical wires, fluorescent lights Essentially impossible to avoid Amplitude: 1 – 3x EMG signal Motion artifact: 0 – 20 Hz Electrode/skin interface, electrode cable Reducible by proper circuitry and set-up Inherent instability of signal Amplitude is somewhat random in nature Frequency range of 0 – 20 Hz is especially unstable > removal

Maximizing quality of EMG signal Signal-to-noise ratio Highest amount of information from EMG signal as possible Minimum amount of noise contamination As minimal distortion of EMG signal as possible No unnecessary filtering No distortion of signal peaks No notch filters recommended, e.g. 60 Hz Differential amplification Reduces electromagnetic radiation noise Dual electrodes Electrode stability Time for chemical reaction to stabilize Important factors: electrode movement, perspiration, humidity changes Improved quality of electrodes Less need for skin abrasion, hair removal

Nerve conduction measurement motor nerve conduction study, MNCS) sensory nerve conduction study, SNCS)