Neuroprosthetics and Brain-Computer Interfaces

By Ellery Buntel

An introduction to neuroprosthetics

As our understanding of, and our ability to connect with, our brains increases, many researchers have set about adapting new technologies to the goal of creating prosthetics controlled by the brain itself. These researchers have created working limbs, games, robots, [1] and even machines that can connect brains together [2]. Some of these so called neuroprosthetics are used by people who suffer from diseases like ALS that prevent them from interacting with normal control schemes, and others are meant to provide a wholly new method of interacting with computers.

These systems most commonly use non-invasive neuroimaging methods like EEG to measure their user’s brainwaves, and then act according to the contents of those signals. Different neuroprosthetics measure different parts of the brain, and utilize different neuroimaging methods to do so. There are also invasive neuroimaging methods that can provide better control to users, but require significant surgery to be used. Modern neuroprosthetics, using these brain-imaging techniques, allow users to control limbs, write, and control computers solely with their brains.

This is an example of one such brain-imaging system

However, despite the extraordinary progress that has been made in neuroprosthetics, there are many technical challenges that still need to be overcome before they can truly become an effective replacement for other control schemes.

What are the challenges associated with neuroprosthetics?

Despite the incredible promise shown by many prototypes created by researchers in the field, there remain many serious difficulties that prevent neuroprosthetics from becoming mainstream.

1. Perhaps the most serious challenge that needs to be overcome in order to develop truly useful neuroprosthetics is that of efficiency. The brain is, by far, the most complex machine mankind has yet encountered, and for neuroprosthetics to effectively replace other modes of interaction and communication they need to interface with the brain on an amazingly complex level. Many brilliant researchers have successfully created interfaces with the brain, but these interfaces are almost always far less efficient than normal communication and interaction methods. For example, spellers are a common neuroprosthetic which allow users to spell out words using only their brain. These spellers are a blessing for users who have no other way of communicating, but they remain far slower and more difficult to use than traditional communication methods [3]. In order for neuroprosthetics like spellers to become viable replacements for normal communication methods, they must become much more efficient.
2. Another common issue with neural control methods is training. Many — if not most — BCI systems require significant training on the part of the user to be able to effectively control them. For most people, this training is doable, but it is not uncommon for some users to be unable to utilize BCI systems entirely [2,4]. A better understanding of how to train users to control these systems will certainly be required if we want to see neuroprosthetics openly available.
3. Another common issue faced mainly by users of neuroprosthetic limbs is the lack of haptic feedback. Essentially, limbs are significantly harder to use without the feeling of touch [2]. In particular, it is very difficult to moderate the strength of your grip without the sensation of touch. Researchers have attempted to remedy this, but it is likely that it will always be harder to control a limb that doesn’t have the ability to feel.

This is one example of a functioning neuroprosthetic limb

These difficulties have proven to be a significant obstacle to the creation of sophisticated neuroprosthetics. However, even these monumental challenges are being taken on by researchers in the field, and as these technologies mature it is likely that we will find ways to work around or overcome these problems all together.

The future of neuroprosthetics

Scientists today are applying state-of-the-art advances in deep learning, neuroimaging, and robotics to try to make neuroprosthetics more viable for the average user. One team achieved an unheard of level of accuracy when classifying speech from users brainwaves [4]. This breakthrough could become the foundation for a new and more reliable method of controlling neuroprosthetics. Another team has successfully proved the feasibility of restoring tactile feedback to human amputees [1]. Yet another team has created a robot that could feel based on a reproduction of a rat’s whiskers [1]. These breakthroughs could well result in artificial limbs that feel just like a biological limb. These are just a few recent advances in neuroprosthetics. There are many more researchers working hard to make usable neuroprosthetics that will fundamentally change the way people with physical disabilities live. With the help of neuroprosthetics, we will altogether change how people live with disabilities, and how we all interact with the world.

References

1. A. Cangelosi and S. Invitto, “Human-Robot Interaction and Neuroprosthetics: A review of new technologies.,” in IEEE Consumer Electronics Magazine, vol. 6, no. 3, pp. 24–33, July 2017.
2. Jiang, L., Stocco, A., Losey, D. M., Abernethy, J. A., Prat, C. S., & Rao, R. P. (2019). BrainNet: a multi-person brain-to-brain interface for direct collaboration between brains. Scientific reports, 9(1), 1–11.
3. Rezeika A, Benda M, Stawicki P, Gembler F, Saboor A, Volosyak I. Brain-Computer Interface Spellers: A Review. Brain Sci. 2018;8(4):57. Published 2018 Mar 30. doi:10.3390/brainsci8040057
4. Makin, J.G., Moses, D.A. & Chang, E.F. Machine translation of cortical activity to text with an encoder–decoder framework. Nat Neurosci (2020). https://doi-org.ezpxy-web-p-u01.wpi.edu/10.1038/s41593-020-0608-8