Brain Computer Interfaces (BCIs) are some of the most exciting technologies today. These devices enable a direct connection between the brain and computers. BCIs have proven to be invaluable in the medical field, assisting patients with conditions like locked-in syndrome to communicate again.
The capacity to control computers using thought alone is captivating, and I eagerly anticipate their application in gaming soon. For now, I celebrate every breakthrough in the field, including a recent neural implant mentioned by Ars Technica, which allows users to speak at near-natural speeds.
Before his passing in 2018, Stephen Hawking was notable not just for his scientific achievements but also for his distinctive computer-generated voice. Many may not realize that his system required significant time to translate thoughts into speech, as it relied on muscle movements to select letters on a screen. This process typically took him about a minute to verbalize a single word.
This innovative technology being developed by neuroprosthetics researchers at UC Davis enhances communication by linking a neural prosthesis directly to the brain and translates brain signals into sounds directly instead of breaking down words into letters. This approach allows for a more fluid and natural speaking style and is significantly quicker than traditional methods.
The initial tests of this technology required the implantation of 256 microelectrodes into a specific brain area responsible for vocal tract muscle control. This information is fed into a neural decoder driven by an AI algorithm. This machine is designed to interpret brain activity in a more nuanced manner, enabling variations such as pitch shifts for questions or conversational interjections like “hmm.” The most exciting aspect is that this process occurs almost instantaneously, with a latency around 10 milliseconds, creating a seamless experience.
Although it has its limitations, the promising results observed so far suggest that a patient transitioned from barely understandable outputs to engaging in complete scripted conversations that others could comprehend. Communicating unscripted, they achieved understanding roughly half the time — a monumental improvement from prior efforts.
The next steps involve refining the technology, as 256 electrodes are relatively few for such intricate tasks. For example, other systems, like one developed by the co-founder of Neuralink, utilize 4,096 electrodes. While these non-invasive options are promising, they may face different challenges due to their distance from the targeted signals.
With the ultimate goal of restoring the ability to both speak and exercise control for those in need, every development in this field is celebrated. I look forward to the day when such technology becomes commonplace in gaming, allowing us to navigate dialog options with just our thoughts.
