How BCI works
Brain Computer Interfaces (BCI) are systems that allow brain waves to be directly measured and used to control a device.
Brain waves are generated by activity in brain cells (neurons) and can be measured with high or low spatial and temporal resolution depending on the measurement method.
Biofeedback sensors and signal processing systems allow those signals to be recorded and converted into usable actions to manipulate flow control in intelligently designed machines.
Neurons
Neurons are signal generators in the brain which function as multi-input single-output logic gates. The inputs are known as dendrites, while the outputs are known as axons. When a neuron is stimulated it becomes excited and generates an action potential. The action potential propagates along the axon to stimulate the dendrites of the next neuron.
Action Potentials
Action potentials are electrical signals generated by polarized chemicals flowing in and out of the neuron through the plasma membrane via ion pumps.
Radiation
Signals are transmitted from the neuron to the scalp through electrical radiation. As the signal passes through more tissue, the temporal resolution (bandwidth) of the signal degrades. The common modes of signal acquisition are
- Electroencephalography (EEG) - which uses electrodes placed on the scalp.
- Electrocorticography (ECoG) - which uses electrodes placed on the surface of the brain.
- Local Field Potential & Single Units (LFP & SU) - which uses electrode arrays that are implanted into the brain.
Homunculus
The motor cortex is a transverse strip crossing from ear to ear along the top of the brain. Individual regions of the cortex correspond to active control of specific body parts and can be identified using the homunculus map. The somatosensory cortex runs parallel to the motor cortex and provides information relating to intended control of their associated body parts.
10-20 System
Researchers and scientists have agreed upon a standardized system for defining scalp points and correlating them to brain cortices. It is known as the 10-20 system and defines the points using percentage spacing based on the nasion and inion of the skull, as well as, the ears.
Amplifiers
In purely electrical systems, a common ground is used between systems to record analog signals. For bioelectric signals, a common ground is not possible, so special amplifiers must be used. The special amplifiers include differential amplifiers with high signal to noise ratios, as well as, right-leg driver (DRL) that creates a ground bias current.
Digitizing, Aliasing, & Filtering
Once the signal has been properly amplified, it is filtered and digitized using analog-to-digital converters (ADCs). The ADC takes discrete measurements, or samples, of the analog signals amplitude at a regular rate, or frequency, so that it can be recorded and reconstructed.
If the ADC sample frequency is slower than the signal frequency, the signal will be reconstructed to look like a lower frequency signal. This process is called 'aliasing' and can cause errors in data collection and acquisition. This concept is known as the Nyquist frequency.
Aliasing can be prevented using two steps. First, the sample frequency must be greater than twice the maximum frequency of your signal. And second, lowpass analog filters can be used, in front of the ADC, to eliminate frequencies above the desired sample frequency, from being aliased into the desired bandwidth.