Where are we with brain-computer interfaces?

Your brain is crackling with the activity generated by roughly 100 billion neurons and their connections. This electrochemical energy controls your subconscious bodily functions, like your heartbeat and breathing, as well as your conscious thoughts and movements.

Brain-computer interfaces (BCIs) tap into those signals, acting as a bridge between the brain and the external world to help people communicate, control a physical extension of themselves or use a smartphone.

There is growing excitement, and not a little hype, around the BCI research field and how to shape its future. Facilitating assistive devices through competition is one of the CYBATHLON disciplines, so what is the current state of development with this technology?

How do BCIs work?

BCIs work by using electrodes to pick up brain activity and transmit these electrical impulses down a wire to a computer. To do this, the electrodes may be placed individually or in a cap directly onto the outside of the skull (non-invasive), attached to the surface of the brain through surgery (minimally-invasive) or implanted into the brain (invasive).

The closer or deeper into the brain the electrodes are, the better the contact and reception for picking up brain signals and therefore the better level of control in the device. The trade off is in the scale of surgery required.

The signals are then transferred into a computer and, using mathematical models or other tools, turned into an output, which could be used to communicate or control something, like a prosthetic hand.

What applications are BCIs currently being developed for?

There are two main areas that BCIs are being developed for: clinical or non-clinical applications.

Clinical applications range from helping patients who have lost the ability to communicate, to those who may be unable to move or who have lost a limb and, with it, a vital movement function.

Non-clinical – or lifestyle uses – could involve using BCIs to avoid distraction when driving, improve learning or working environments, or allow users to control their phone or digital book reader with their brain rather than their fingers. In each case though, the BCIs tap into the users’ brain signals and learn what they want to say or do.

What are the current challenges for BCIs?

One of the chief challenges lies with the electrodes used to record the brain signals.

If you implant an electrode into the brain, you only want to do it once because it is invasive surgery. This means they need to be made of materials that are safe and long-lasting inside the human body.

Currently that varies. In some cases, the electrodes can continue to work for years, but in others they degrade very quickly. Materials scientists are working on developing biocompatible electrodes that can last reliably inside the brain without degrading or being rejected by the body.

The electrodes are also connected to computers by cables that go through the skull, a cumbersome set up and a potential source of infection. Developing wireless communication from electrodes to the computer will remove the need for cables, provided the circuits can be kept cooler than 40℃ so as not to harm the brain.

Professor Gernot Müller-Putz, Head of the Institute of Neural Engineering and Laboratory of Brain-Computer Interfaces at Graz University of Technology, and his group have been working on developing just such a system in a large European project. It will be primarily for people who have lost speech, for example in people who have amyotrophic lateral sclerosis (ALS), the most common form of Motor Neuron Disease (MND).

Non-invasive BCIs, while not requiring surgery, will also still need to become as unobtrusive as possible before being widely accepted.

“I don’t think that people walking around wearing electrode caps with cables attached to a computer will be accepted for people to adopt a BCI to control their smartphone or another device. We need to be able to integrate this technology into your glasses or earphones,” says Professor Moritz Grosse-Wentrup, Head of the Neuroinformatics group at the University of Vienna.

Then there’s the computing power and algorithms needed. BCIs and users need to train each other; the BCI learns what signals are coming from the brain and the user learns how they control the BCI. This requires continuous calibration and feedback because so many variables – including shifting electrode positions, tiredness or stress – can affect how the signals are interpreted.

The article is written by Edward Brydon