WHAT IS A BRAIN-COMPUTER INTERFACE (BCI)?


A brain computer interface is a direct communication pathway between the brain and an external device. BCIs are often aimed at assisting, augmenting or repairing human cognitive or sensory-motor functions. There are two basic type of BCIs, input systems and output systems. An input brain computer interface artificially creates an internal perception, allowing somebody to see, hear, or feel a sensation they have lost, while on the other hand, an output brain-computer interface "gives their user a communication and control channel that does not depend on the normal output channels of peripheral nerves and muscles," potentially providing somebody with a loss of motor control a newfound interaction with their environment. This wikispace will primarily deal with the latter type of BCI.



BRIEF REVIEW
In order to understand brain-computer interfaces further, it is important that there is a general understanding and review of several concepts.


COMMUNICATION IN THE BRAIN- Neuron to Neuron

Neurons are at the foundation of everything involving the use of our brain. They are involved in what we think, feel, experience and even remember. Formally, a neuron is "an electrically excitable cell that processes and transmits information by electrical signaling." Information is relayed from neuron to neuron through various neuronal networks in the brain via these electric signals, which are generated by differences in electric potentials on a cell's membrane, brought about by the movement of ions into and of the cell. At rest, a neuron is passively leaking potassium(K+), creating a resting membrane potential, or difference in voltage between the interior and exterior of the cell. When a neuron becomes excited, there is an influx of sodium(Na+). Once a certain voltage level, or threshold is reached,
Neuron to Neuron Communication
Neuron to Neuron Communication
all sodium channels will open, resulting in an exponential increase in the amount of influx of sodium into the cell. This ultimately causes its depolarization. Once a certain level of depolarization is reached, K+ channels are triggered open and there is an eflux of potassium from the cell, resulting in its re-hyperpolarization. This short lived rapid rise and fall of the electric membrane potential is called an action potential. An action potential is essentially the electric signal that neurons "communicate" with. Once generated in the axon hillock, an action potential is propagated down the neuron's axon until it reaches the pre-synaptic terminal where is eventually causes depolarization and generation of an action potential in the connecting neuron. It is by this means that information is transfered from location to location in the brain.


CONTROL OF VOLUNTARY MOVEMENT- The Brain

How is the information sent along neurons processed through the different parts of the brain? That is, how exactly is it that we create a movement? In reaching for a glass, how does our hand and arm know what to do? The human brain is the main force behind our movements. Many different regions of the brain are involved in movement control, but for the purpose of learning about brain computer interfaces, we will primary focus on the motor cortex. The motor cortex is made up of three different regions: the premotor area(PMA) and the supplementary motor area(SMA) both of which are located
Creation of Movement
Creation of Movement
in area 6 and the primary motor cortex, which is located in area 4. The premotor area is primarily responsible for integrating sensory information that helps to guide movements of the body. It is also responsible for controlling the proximal muscles. The supplementary motor
Descending Pathways
Descending Pathways
area on the contrary, aids in coordinating movements of both hands and in planning complex movements. Lastly, the primary motor cortex is responsible for controlling muscles on the contralateral side of the body. When creating a movement, the frontal lobe is involved in the planning of it. Information regarding such things as current body position is received as input by the frontal lobe, which then transfers the commands it generates to area 6 of the brain. This region is responsible for deciding which set of muscles to contract. Finally, area 6 transfers this information to area 4, which is responsible for activating specific muscles or groups of muscles. The premotor cortex is therefore involved in the actual activation of the muscles.

CONTROL OF VOLUNTARY MOVEMENT- Descending Pathways

How is it that information is conveyed from the primary motor cortex to the muscles themselves? Information from the cortex regarding voluntary movement is relayed to the brainstem and the spinal cord via two main descending tracts. The first is the corticobulbar tract. This tract is responsible for cortical control of fascial muscles. The second and most important tract is the corticospinal tract. This tract connects to motor neurons in the spinal cord. Stimulated motor neurons from the spinal cord then innervate various axial muscles along with the various extremities of the body. The corticospinal tract can be further subdivided into two tracts referred to as the lateral system and the ventromedial system. Fibers in the lateral system are responsible for controlling muscles of the contralateral extremities as a result of decussation of the descending fibers in the medulla. The ventromedial system on the other hand have fibers that do not cross and are responsible for controlling the body's axial muscles.

For further explanation of movement and the brain please refer to this site.




OUTPUT BRAIN-COMPUTER INTERFACE


OVERVIEW OF AN OUTPUT BCI

Output brain-computer interfaces "give their users communication and control channels that do not depend on the normal output channels of peripheral nerves and muscles". Essentially, this device is able to "read a person's thoughts" by decoding the brain waves that a person is emitting. These signals are then created into commands that an external device, such as a computer or prosthetic limb can then interpret and act off of. How exactly is it though that a machine can read somebody's thoughts? As reviewed above, communication between neurons is the way in which our brain communicates with itself and the rest of our body. As a result, every thought we have is a result of neurons communicating with each other via action potentials,i.e. electric signals. As these action
external image brain-computer-interface-2.gifpotentials are propagated down a myelinated (remember: myelin essentially acts as insulation) axon, some of the electric signal escapes. It is the job of a BCI to detect these escaped signals, interpret them, and then transfer them into commands that a device can understand. These brain signals are often captured by electrodes that are either attached to the surface of the scalp, implanted directly into the gray matter of the brain itself or are attached directly beneath the skull. Once electrodes acquire a signal, it is then amplified, filtered and translated by the computer or machine such that commands can be carried out that reflect the intentions of the user. For example, someone who is a tetrapalegic is unable to move their extremities, and as a result, incapable of moving the cursor on a computer screen. A BCI would allow such a patient to again be able to interact with a computer. While using a BCI, if a person thinks about moving the mouse, electrodes in the brain pick up on the brain signals, and sends them to a computer which will interpret what the intended action of the person is, and then perform that action. As a result, a patient could literally move a mouse just by thinking about its movement on the screen.

BRAIN-MACHINE LEARNING

In order to be able to interpret the action a patient wants to perform, there is often a certain amount of "learning" that both the user and the BCI must undergo. The user of the BCI must learn to modulate brain waves such that they are being used in a way that maximizes BCI performance and/or the device itself, must "learn" to identify, interpret and adapt to the most salient neural signals that best decode the intended action of the user. Essentially, a BCI must be equilibrated with the thoughts and specific brain waves of its user.

There are two specific ways for a BCI to be equilibrated with the user's brain waves, either by an open-loop paradigm or a closed loop paradigm. In an open-loop paradigm, the computer is altered to be able to better interpret brain activity. So while performing several trials during initial training with the BCI system, the subject is unaware of the way in which the computer is interacting with their recorded brain activity. After completion of the trials, a specific computer task will be matched with the recorded activity. So, for example, brain activity of the user would be recorded while performing several trials in which the subject is thinking about moving a cursor to the left. After the trials and recordings had been completed, somebody would manually correlate a desired computer task(i.e. leftward cursor movement), with the recorded brain waves. A closed loop paradigm on the other hand is where brain activity is modified so it can be more representative of computer activity. A subject is therefore provided with real-time feedback of how their brain activity is performing. That is, there is an optimal level of brain activity that must be reached in order for optimal computer functioning to occur. The user is therefore an active participate in learning how to best control the computer through their brain activity. The closed loop paradigm is preferential over the open-loop paradigm,

COMPONENTS OF BRAIN COMPUTER INTERFACE- in further detail


BCI.jpgThere are several elements that are important in order to have a functioning BCI. The first important aspect of a brain-computer interface is adequate signal acquisition. That is, the system has to be able to receive brain signals or receive information input. It then must be able to convert this raw information into a useful device command. This is known as signal processing. Within signal processing, there are two essential elements. The first is that of feature extraction. It is vital that the device can detect the difference between important and superfluous information retained during signal acquisition. It is then necessary that this information be converted into device commands. This is known as translation. It is important that during this stage of signal processing that a relationship be determined between an electrophysiological event and a specific motor or cognitive task. For example, if we are working with a tetrapalegic attempting to move a cursor on a computer screen, the BCI must first determine and recognize the change in the electrical signals being emitted while the patient is attempting to move the cursor ( feature extraction). It must then be able to attribute that specific variation in electrical signal to a specific movement of the cursor (translation). The actual movement of the cursor would be the device output. That is, device output is "the overt command or control functions administered by the BCI system". Device output can include a conglomerate of results, ranging from the cursor example above to the movement of a robotic arm. The fourth and final component of a brain computer interface is that of operating protocol. This refers to the way in which the user interacts and controls the BCI system. "The 'how' includes such things as turning the system on or off, controlling what kind of feedback is provided and how fast, the speed with which the system implements commands, and switching between various device outputs." It appears as these parameters are vital in the actual functioning of the brain-computer interface.

PLACEMENT OF ELECTRODES-in further detail

During signal acquisition, the electrophysiological signals being emitted from the brain are usually recorded with the aid of electrodes. Electrodes are essentially electrical conductors responsible for making contact with the nonmetallic part of a circuit. In this case, the brain. Varying from where the neural signals are received, BCIs can classified as to whether they use electrodes that are non-invasive, invasive, or partially invasive. Electroencephalography(EEG)-based BCIs obtain electrical signals from the scalp. This has been the most prominent way in which
BCI2.jpg
Layers in Which Electrodes are Placed
non-invasive BCIs receive signals. This is a result of the relatively safe and practical application of this type of brain computer interface. This type of BCI is typically used for patients with locked-in syndrome, a condition where sensory or cognitive functions are normally in tact, but there is complete destruction of the peripheral and central motor system. As a result, a person becomes completely paralyzed. Using an EEG based system, a locked-in-patient will normally wear a head set and be face to face with a computer screen that presents them with various words or letters. When a desired letter appears, the patient "selects" it essentially by just thinking about it. The BCI is then able to determine that a letter has been chosen as a result of the patient's change in brain waves. It is through this sort of a mechanism that someone in a locked in-state is able to communicate with the outside world (see video below!) There are however significant drawbacks to using an EEG-based system. It has been found that in order to detect large enough signals for use by the system, a large brain area must be involved. This results in EEGs having poor spatial resolution. In addition, the signals from EEG devices do not contain information about certain parameters of movement such as position and velocity, and there is also risk of interference from the electrical activity of additional cranial musculature. This ultimately results is signal distortion. Additionally, in order to be able to use an EEG system you face other obstacles such as "long training periods, noisy signals, a need for continuous professional attention, slow spelling speed, electrode and skin problems with long recording times, and the need for controlled attention focus during spelling mistakes."

Invasive BCIs, on the contrary, appear to provide us with better alternatives as well as more informative signal reception. There are two basic type of invasive BCIs. They include "single units" microelectrodes, which transmit activity of the action potentials of individual neurons, and field potentials, which monitor brain activity from within the parenchyma. Hypothetically speaking, one would think that embedding electrodes into the cortical layers so that they transmit signals from single neurons, would be the most efficient way to monitor brain signals. This does appear to be the case, but however, for only a short period of time. There appear to be two main problems that one faces when dealing with invasive BCIs. First, penetration of the parenchyma by electrodes causes vascular and localized neural damage, and also exposes the patient to potential central nervous system infections. Secondly, the existence of an alien substance in the brain, causes the electrode to become encapsulated in scar-tissue over time. This can result in disruption of the signal. As a result of the significant disadvantages of the past two types of BCIs, a third type has emerged within the past few year as the dominant type of BCI signal platform. It is that of an electrocorticography-based system (ECoG). An ECoG based system records electrical activity of the brain beneath the skull. Unlike the invasive BCIs, since ECoG electrodes are not implanted in the parenchyma, they cause less damage to the brain and don't experience as much signal deterioration from encapsulating scar tissue. Also, as opposed to EEG, ECoG can detect signals better. They can pick up on high frequency gamma wave activity produced from cortical neurons. As a result of these advantages, electrocorticography-based BCIs have become viewed as the best of both worlds and the preferred form of system. In addition to the above mentioned benefits, when again dealing with patients with locked-in syndrome, partially invasive BCIs give people a much faster ability to select letters. Nevertheless, many patients still voice preference for noninvasive BCIs regardless of the long training period time and slow selection rate of letters. From such a person's perspective "time is not an issue if one is completely paralyzed".



BCIs, THE MOTOR CORTEX, and FUTURE RESEARCH
Most of the BCI platforms on the market today are derived from brain signal changes that occur in the motor cortex of the brain. That is, invasive output BCIs place electrodes in or near the primary motor cortex as a result of this regions ability to control motion of the contralateral extremities. Most people benefitting from BCIs therefore are people whose motor cortex remains intact, i.e. people suffering from spinal cord injuries(SPI), neuromuscular disorders or amputations. This however, leaves a significant portion of motor impaired people without aid. Specifically, patients suffering from unilateral strokes are unable to take advantage of current BCI systems due to the damage in one of the hemispheres of their cortex. There are mass research efforts therefore being employed in an attempt to find ways to restore some form of motor abilities for these patients. Current research includes analyzing areas of the brain responsible for ipsilateral movement of the bodies and employing other areas of the brain, such as the sensorimotor cortex, as possible locations from which to receive and employ brain signals for interpretation and use by a BCI.

SUMMARY

Overall, we have barely scratched the surface of brain-computer interfaces. There is so much underlying potential, that if discovered, could change the lives of hundreds of thousands. We have already begun helping the lives of some, and I believe it is only a matter of time before we can create a system that allows a stroke victim to move his or her limbs again. We are living in a time at the forefront of technology, in which sci-fi of the past is beginning to turn into reality. As we delve deeper in our understanding of the brain, our emerging technology will become unparalleled to anything we have ever created.








GLOSSARY

Brain Computer Interface: a direct communication pathway between the brain and an external device.
Signal Acquisition: the BCI system's recorded brain signal or information input
Signal Processing: the conversion of raw information into a useful deice command
Device Output: the overt command or control functions administered by the BCI system
Operating Protocol: the manner in which the system is altered and turned on and off
Electroencephalography(EEG)-Based BCIs: non-invasive BCI which obtains electrical signals from the scalp
"Single Units" Microelectrodes: invasive BCI platform that transmits activity of the action potentials of individual neurons
Field Potentials: invasive BCI platforms which monitor brain activity from within the parenchyma
Electrocorticography (ECoG)-Based BCIs: A partially invasive BCI platform which records electrical activity of the brain beneath the skull
Primary Motor Cortex: Part of the motor cortex responsible for controlling muscles on the contralateral side of the body

SUGGESTED READINGS

HowStuffWorks
Provides a simple and brief overview of the workings of a brain-computer interface.

Mechanics of a BCI
This is a link to a pdf file, detailing the specific mechanical workings of a BCI. This is for those who desire to learn the rudimentary engineering concepts behind such systems.

Brain to Brain!
This is the link to a video capturing an experiment in which two people try to communicate with each other via brain waves through a computer. Pretty neat!


QUIZ

MULTIPLE CHOICE

1. All of the following are important to have a functioning BCI except:
A. signal acquisition
B. signal processing
C. device output
D. operating protocol
E. all of the above
F. none of the above

2. Which of the following is not a type of BCI system
A. EEG-based system
B. ECoG-based system
C. field potential-based system
D. EKG-based system
E. single-unit micro-electrode-based system

3. Which region of the brain is the primary area in which invasive electrodes are placed?
A. pre-motor cortex
B. frontal lobe
C. primary motor cortex
D. supplementary motor areas

4. The descending pathways from the motor cortex responsible for voluntary movement are,
A. corticobulbar and anterolateral tracts
B. corticobulbar and vestibulospinal tracts
C. corticobulbar and corticospinal tracts
D. ventramedial tract and lateral systems
E. cortiospinal and tectospinal tract
F. A and D
G. C and D

5. Which of the following are considered invasive BCIs?
A. EEG-based system
B. ECoG-based system
C. field potential-based system
D. single-unit micro-electrode-based system
E. B and D
F. A, B and E

TRUE OR FALSE
6. The motor cortex is made up of two regions.
7. An electroencephalography is an invasive BCI
8. BCI stands for brain-computer interface

SHORT ANSWER
9. Describe briefly how a voluntary movement is initiated.
10. Describe which type of BCI is the best and why.
11. Describe the main difference between an open-loop and closed-loop paradigm.

answers
1. E
2. D
3. C
4. G
5. E
6. F
7. F
8. T
9. When creating a movement, the frontal lobe is involved in the planning for any given movement. The frontal lobe receives input from many other regions of the brain about such things as current body position. The frontal lobe then transfers its commands and plans to area 6 of the brain which is responsible for deciding which set of muscles to contract. Finally, area 6(premotor cortex)transfers this information to area 4, which is responsible to activating specific muscles or groups of muscles.

10. ECoG: Unlike the invasive BCIs, since ECoG electrodes are not implanted in the parenchyma, they cause less damage to the brain and don't experience as much signal deterioration from encapsulating scar tissue. Also, as opposed to EEG, ECoG can detect signals better. They can pick up on high frequency gamma wave activity produced from cortical neurons

11. In an open-loop paradigm, the computer is altered to be able to better interpret brain activity, while closed loop paradigm on the other hand is where brain activity is modified so it can be more representative of computer activity



REFERENCES

Kandell, ER., JH. Schwartz, and TM. Jessell, eds. Principles of Neural Science. 4th ed. New York City: McGraw-Hill Companies, Inc., 2000.

Lee,U., Lee H.-J., Kim S., Shin H.-C. Development of Intracranial brain-computer interface systems using non-motor brain area for series of motor functions. Electronics Letters 42: 2006

Leuthardt, Eric C., Schalk, Gerwin, Roland,Jarod, Rouse,Adam, Moran, Daniel,W. Evolution of brain-computer interfaces:going beyond classic motor physiology. Neurosurg Focus 27: 2009.

Kotchetkov,Ivan S., Hwang, Brian Y.,Appelboom,Geoffrey,Kellner,Christopher,Connolly,Sander. Brain-computer interfaces: military, neurosurgical and ethical perspective. Neurosurg Focus, 28: 2010.

Nicolelis, Miguel A.L., Lebedev, Mikhail. Principles of neural ensemble physiology underlying the operation of brain-machine interfaces. Nature Reviews,10: 2009.

From You Brain to the Machine. Mechanical Engineering: 2010.

Leuthardt. Engineering new treatments for stroke with brain-computer interfaces. Future Neurology: 133-137, 2009.

Birbaumer, Niels. Breaking the silence: Brain-computer interfaces(BCI) for communication and motor control. Psychophysiology 43: 517-532, 2006.

Canadian institute of health research. The brain from top to bottom. Retrieved from http://thebrain.mcgill.ca/flash/d/d_06/d_06_cr/d_06_cr_mou/d_06_cr_mou.html#4

Grabianowski,Ed. How brain-computer interfaces work. Retrieved from http://computer.howstuffworks.com/brain-computer-interface.htm

Karapetyan, Vahagn. The role of brain plasticity in BCIs. Retrieved from http://hmb300neurowiki.intodit.com/page/the-role-of-brain-plasticity-in-bcis