In recent years, developments in brain computer interface technology have been turning science fiction into science fact. Some unbelievable achievements have been made, including giving locked-in syndrome suffers a means to communicate, allowing amputees to feel their prosthetic limbs and restoring sight to the blind. And in 2013, Harvard researchers made a rat’s tail wiggle with only the power of their minds! The brain computer interface is revolutionising medicine, technology and even gaming. But some of the current research may make people feel a little uneasy….
The brain computer interface (BCI) is a system that allows a computer to read human brain activity and interpret the signal, as well as inputting new signals back into the brain. Essentially, the BCI allows computers to read your mind. Well, sort of. The Brain-computer interface has had an incredible impact on the quality of life for suffers of paralysis, Amyotrophic lateral schlerosis (ALS), myopathy, spino-cerebellar ataxia (SCA), cerebal palsy, and is even now being adapted to help autistic children train their minds and improve concentration. BCI can allow patients with limited mobility to control motorised wheelchairs and prosthetic limbs, and communicate with the world. It has also been used to identify signs of life in patients with no means to communicate. More recently, it has been applied to a far wider range of uses, including video games and fashion.
How Does BCI Work?
Your brain contains about 100 billion neurones, connected in an imaginably complex network. Neurones communicate with each other and the rest of the body via electrical impulses, generated by ion imbalances across the neuronal membranes. Just like electrical cables, neurones must be insulated to prevent the electricity escaping, so each neurone is coated in a substance called myelin, which is pretty good at insulating the signal. Pretty good, but not amazing – some of the electrical impulses escape into the surrounding tissue and it is these escaped impulses that are read by computers in the brain computer interface. Also key to BCI is the amazing flexibility of the human mind – neurones are highly adaptable and can alter their connections with training. This is how patients with brain damage can regain lost function; they train other parts of their brain to perform the jobs of the damaged regions. Likewise, we can train our brains to produce readable outputs, and interpret new inputs.
The most common way of measuring brain waves for BCI is using Electroencephalography (EEG), whereby multiple electrodes are placed on the outside of the skull and detect escaped electrical impulses over a matter of minutes. Another method, Electrocorticography (ECoG), works in exactly the same way, but is able to access a higher quality signal because the electrodes are placed inside the skull. EEG and ECoG can be used to detect the thoughts of a participant directly by detecting electrical impulses produced by a particular region of the brain (Event-related potentials, ERP), or indirectly by measuring desynchronisation between active and inactive neurones (Event-related desynchronisation, ERD). A third method (Steady-state evoked potential, SSEP) involves detecting electrical impulses produced by the brain in response to sensory input such as light or sound. The majority of BCI devices work by detecting the whispers of electrical activity that escape the myelin sheath, but there is more than one way to skin a cat. Other methods, such as Magneto encephalography (MEG) and Functional Magnetic Resonance Imaging (fMRI), work by detecting changes in blood flow or magnetic fields in the brain. These are both indirect ways of measuring changes in brain activity – blood flow increases in active neurones, and electrical impulses produced by those neurones produce magnetic fields.
So, we can detect electrical impulses in the brain, directly or indirectly, and use those to interpret what the brain is doing. It sounds like a form of mind reading, but BCI always requires conscious effort on the part of the participant. Participants must train their brains to produce the correct impulses to send the right message to the computer. This might involve thinking about a particular movement, or directing your attention to the appropriate light source, but the participant must learn to speak to the computer in the right way. This means that it takes time to master BCI – depending on the device and the signal being measured, it can take 30 hours or more to learn to control a BCI. It also means that the quality of the signal is influenced by the condition of the participant – controlling a BCI is tiring, and fatigue has been found to negatively affect signal quality. Finally, it means that, at least at present, BCI cannot be used to read somebody’s mind without consent.
Early applications of BCI technology allowed paralysed patients to communicate with the outside world using technology such as the P300 speller, which reads event-related potentials to allow patients to spell out words and sentences. At present, the accuracy of these devices is still relatively poor compared to communication systems like the one used by Stephen Hawking, which relies on minute facial movements. However, for sufferers of locked-in syndrome, BCI is the best option for communication. Although the technology available commercially remains in its infancy, research promises great things for these communication devices, and earlier this year, researchers from East Tennessee State University announced in the journal Science that they had successfully enabled a locked-in patient, who had suffered a brainstem stroke and total paralysis, to spell using a BCI-based technology. Over 13 months the patient was able to learn to communicate with the computer using only his mind, to spell out words and to communicate with the world again.
One of the major goals of BCI has been to enable the control of external objects, for example wheelchairs and robotic limbs. For many sufferers of paralysis and muscular or neurodegenerative diseases, this is means mobility and autonomy, and in recent years, scientists have succeeded in producing mind-controlled artificial hands, legs and arms which can be controlled by the brain. BCI-based wheelchairs also appear to be on the cards; a study published in January this year reported that 15 participants had successfully navigated a BCI wheelchair around a building using event-related potentials.
The applications of BCI extend well beyond medicine. In particular, the multi-million dollar industry of gaming seems an obvious candidate. In 2009, a company called Uncle Milton released Star Wars Force Trainer, where players could control a ball with nothing but the power of their minds. Now, several companies offer BCI headsets for gaming, including Emotiv and NeuroSky. Recently, a Japanese company called Neurowear developed BCI wearable cat ears, which move according to your mood and concentration*. And just the other week, I sat in a pub in North London and used my brainwaves to control a tug-of-war game on a computer screen. You can now buy your very own BCI device for just £30, or you can even make your own!
* It’s not clear exactly how these BCI devices measure ‘concentration’ or ‘mood’, but both alpha and delta brainwaves have been found to alter according to concentration and relaxation levels, so these commercial BCI devices may be targeting those.
Perhaps more interesting than merely reading a mind is inputting new information. Although this sounds conceptually far more difficult, the first ever neuroprosthetic device provided new sensory input to the brain in the 1980s – the cochlea implant. Other sensory inputs have proved more difficult, and 30 years after the cochlear implant began restoring hearing to the deaf, we are only just beginning to be able to tackle more difficult tasks, such restoring sight to the blind, and providing sensory feedback from artificial limbs. Last year, eight people suffering from the degenerative disease retinitis pigmentosa received retinal implants, which enabled them to see some light again. Also in 2013, researchers at the University of Chicago released details of their touch-sensitive prosthetic limbs that can communicate directly with the brain.
This year, amputees in the US and the UK received mind-controlled, touch-sensitive limbs, which surgically tap into the wearer’s own nervous system. Researchers have also been developing sensitive artificial skin that could be used to treat burn victims or cover artificial limbs. One artificial skin, unveiled by researchers at Stanford University this year, is 1000 times more sensitive than real human skin! The skin is made of two layers of self-healing rubber, which sandwich electrodes between them. Acting a bit like a spring, the compression of the rubber creates an electrical charge, which is carried away by the electrodes (just like the nerves in real skin). Although it is only 1mm thick, the artificial skin is so sensitive it can detect the pressure exerted by a single fly landing on it. Meanwhile, researchers in Korea this year announced a new ‘smart skin’ made of silicon that can sense temperature and humidity, as well as pressure. This skin is still in the early stages of research, and has so far only been tested in rats, but the results are promising. Touch-sensitive prosthetics could change the lives of millions of amputees, although the authors have also identified many useful applications in industrial and commercial settings – imagine how little maintenance a self-healing electrical device, with the ability to sense the world around it, would need!
Adding in new sensory input is one thing, but influencing anything else – movement, thoughts, emotions, intentions – remains imprecise if not impossible. However, the sensorimotor cortex, responsible for processing sensory input and controlling movement, is the best-understood region of the brain and is located close to the surface, making it a good target for BCI. In 2013 researchers at Harvard University succeeded in creating a non-invasive Human Brain – Computer – Rat Brain interface, which enabled a human to consciously control the tail of the rat. This was achieved using transcranial focused ultrasound (FUS), which is a non-invasive method for influencing brain activity in particular regions.
Last month, six students in Washington had their brains read using electroencephalography (EEG), which were then transmitted over the internet to their partner who received them through transcranial magnetic stimulation (TMS). Together, brains connected via the internet, the pair had to play a computer game in which timing was crucial. The transmitting student was shown a trigger and they were responsible for sending the brain signal to the receiver to fire at the right time. Similarly, researchers at Cornell successfully connected the brains of two monkeys last year, and allowed one to control the movement of the other. The team sedated a monkey and implanted 36 electrodes into its spine. These were connected to an implant in the brain of a second monkey, monitoring around 100 motor neurones. This monkey remained awake, and was taught to move a joystick. Finally, the team placed a joystick in the still sedated receptor monkey and the transmitter monkey was able to move the joystick in his hand, too! This sounds remarkable, however in both cases, the signal that was sent was a relatively simple one – stimulating the general region that controls motor behaviour in the right hand and causing it to jerk in the hope that this would be sufficient to move the joystick. Impressive, yes. But the researchers admit this is a long way from fine control of movement, let alone complex thoughts, emotions or desires! We still don’t understand how precise movements are encoded in the brain, and anything more exact is simply out of the question at present.
Beyond Science Fiction
Last month, research published in Nature Communications made a bold and rather bizarre claim – that they could now control gene expression in mice with their minds. Sorry, what? Yep, researchers in Switzerland combined EEG-based BCI with optogenetics (genes activated by light) to create a mind-reading device that when activated would turn on a light inside the mouse’s brain, activating specific genes. Although they stress this study remains merely proof-of-concept at this stage, they suggest this technology could have important medical applications. For example, it could be used to deliver drugs to epileptic patients prior to a seizure, or enable patients with locked-in syndrome to self-medicate.
Optogenetics takes advantage of the fact that nature has already created a system of light-sensitive gene expression. It is important in regulating circadian rhythms such as sleep and waking patterns in many vertebrates, and bacteria have engineered sensitive proteins too. By taking an infrared-sensitive gene from bacteria and inserting into a human kidney cell, the researchers engineered a system whereby infrared light would trigger activation of a protein which would lead to a cascade of molecular changes ultimately resulting in the expression of a particular gene. These cells, alongside a remotely-triggered led, was implanted into the brain of a mouse. Hooking this system up to BCI technology allowed eight volunteers to trigger the led inside the mouse’s brain and ultimately control the expression of specific genes.
The Future of BCI
Developments in the brain computer interface are picking up pace – what is the future of BCI? The mind boggles. Both medical and commercial applications of BCI will benefit from improvements in the quality of signal, whilst reducing invasiveness of the techniques. However, major hurdles remain in our understanding of how the brain works, if we are to improve our ability to read it. The brain computer interface is a remarkable achievement, and testament to just how far neuroscience and technology have already come, but compared to the sophisticated processing possible in the human brain, our BCI technologies are still relatively crude. Crude is good enough, though, for patients for whom BCI has brought new mobility, new autonomy and new communication.
Want to Know More?
- Pu et al Semiconducting polymer nanoparticles as photoacoustic molecular imaging probes in living mice Nature Nanotechnology
- Kim et al (2014) Stretchable silicon nanoribbon electronics for skin prosthesis Nature Communications
- Folcher et al (2014) Mind-controlled transgene expression by a wireless-powered optogenetic designer cell implant Nature Communications
- McCane et al (2014) Brain-computer interface (BCI) evaluation in people with amyotrophic lateral sclerosis Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration
- Rao et al (2014) A Direct Brain-to-Brain Interface in Humans PLOS ONE
- Grau et al (2014) Conscious Brain-to-Brain Communication in Humans Using Non-Invasive Technologies PLOS One.
- Shanechi, Hu & Williams (2014) A cortical–spinal prosthesis for targeted limb movement in paralysed primate avatars Nature Communications
- Sellersm, Ryan & Hauser (2014) Noninvasive brain-computer interface enables communication after brainstem stroke SCIENCE
- Kaufmann, Herweg and Kübler (2014) Toward brain-computer interface based wheelchair control utilizing tactually-evoked event-related potentials Proceedings B
- Stingl et al (2013) Artificial vision with wirelessly powered subretinal electronic implant alpha-IMS Proceedings B
- Tabota et al (2013) Restoring the sense of touch with a prosthetic hand through a brain interface Proc. Natl. Acad. Sci. USA
- Yoo et al (2013) Non-Invasive Brain-to-Brain Interface (BBI): Establishing Functional Links between Two Brains PLOS ONE
- Duann et al (2013) BCI-Controlled Videogame for Cerebral Palsy Children Proceedings of the Fifth International Brain-Computer Interface Meeting
- Amiri, Fazel-Rezai and Asadpour (2013) A Review of Hybrid Brain-Computer Interface Systems Advances in Human-Computer Interaction
- Anupama, Cauvery and Lingaraju (2012) Brain Computer Interface and its Types International Journal of Advances in Engineering & Technology
- Fazel-Rezai et al (2012) P300 brain computer interface: current challenges and emerging trends Frontiers in Neuroengineering