Brain-Computer Interfaces (BCI) are technologies that enable a human user to control an external device solely based on their brain activity and without involving the movement of their muscles or limbs. In other words, the activation of neurons in the brain is read by the device in order to generate control commands.

There are two major classes of BCIs being currently developed. The first class, called “invasive BCI”, measures directly neural activity from the human cortex by introducing “invasive” electrodes into the brain. In such cases, one must drill a hole in the human skull and insert recording probes into the brain’s tissue. One example invasive BCI is the one developed by Elon Musk’s Neuralink.

The second type of BCI technology, called “non-invasive” BCI, attempts to record brain activity from outside the skull, without requiring surgery and being completely harmless for the user.

Non-invasive BCIs usually rely on the recording of electrical fields produces by the brain using an extremely sensitive electrical device called an electro-encephalograph (EEG). Because the electrical activity produced by neurons is tiny, and because the electrical fields have to travel through the bone and skin to reach the recording electrode, the EEG signal is very small and prone to noise. External sources of electricity, such as the mains line or devices, such as a monitor, can produce interference signals that are orders of magnitude larger than the useful signal coming from the brain. Thus, there are two major challenges for non-invasive BCIs: filtering by the bone and skin, and external noise. Let’s discuss them next.

Filtering. You probably know that when the neighbor next door turns on music loudly, the main feature that you hear is the bass, while the high-frequency treble is filtered out by the wall. Pretty much in the same way, the EEG has a difficult time “hearing” the high-frequency oscillations that your brain generates. This is unfortunate, because “fast” communication between the brain and the BCI device requires a lot of information transfer, which requires high-frequencies.

Noise. Since the useful signals coming from the brain are so small, the external noise typically dominates the EEG signal, which makes it very difficult for the BCI to “home in” on the true, neural signals that your brain generates.

Due to filtering and noise, traditional BCI technology can only harvest slow brain waves, which are typically larger and can cross through the bone and skin. This leads to the so-called “bandwidth problem”: the EEG signal is not capable of carrying enough information to enable a fast BCI device. Traditional BCI devices are so slow, that they are not practical for demanding control scenarios, such as in gaming. For example, classical BCIs allow users to make one movement every 2 seconds. Not very useful, is it?

At the Transylvanian Institute of Neuroscience (TINS – https://tins.ro), we developed a patented technology that allows the BCI to capture the fast brain waves that your brain generates, called “gamma waves”. These waves have a frequency of 30-120 Hz, and enable very fast control of the BCI device, with up to 4 commands per second! This performance is unprecedented and unmatched by any other BCI device on the market!