Brain-Computer Interface

Brain-computer interface is a device, often a computer-based system that creates a direct communication pathway between a brain’s electrical activity and an external output. To put it simply, prior to moving your arms or legs, signals are sent to your brain. A BCI on the other hand doesn’t require an intermediary. It reads signals directly from your brain and translates them into commands a computer can comprehend. Let’s see how it works in more detail. Based on a report published by the U.S. Government Accountability Office, lesser than 70 people globally had used an implanted decoding BCI as of September 2024.

It is now an established fact there is an input and output but no intermediary except if you call the device that connects the input and output an intermediary. The signal from input to output is extremely fast that it feels almost continuous.

BCI converts the neurones activity into commands that a machine can comprehend. For that, electrodes are placed on the scalp or closer to the brain to capture the signals produced when groups of neurones fire. Since the initial signals are noisy, the system amplifies them to filter out muscle twitches and other interference to identify patterns that correspond to an action. There will be an “understanding” period whereby the algorithm learns the individual’s unique neural signature that helps the machine-learning model map patterns into digital commands.

The external device are conventionally a computer cursor, a speech synthesiser, a robotic arm or a wheelchair. The user will also receive feedback through different senses including sound, sight or touch which allows the brain to fine-tune its signals and enhance control. There are also non-invasive EEG systems deployed in research and rehabilitation, and implantation provides better precision for individuals with severe paralysis. However, it still requires surgery and there are concerns on long-term stability. ²

^{Image generated by grok}

Morgan Stanley projects $1.5B in cumulative BCI revenue by 2035 without exceeding $500M in annual run rate until 2036, with lesser than 3% of penetration by 2045. Richard L. Wells wrote a piece on Tech Times that Synchron’s $200M Series D raise is specifically meant to finance its 2026 trial, the study the company has to clear before filing for the first-ever FDA premarket approval of an implanted BCI.

On the flip side, GAO is observing the tech’ transition from medical implants to human amplification. This rose as a result of the technology’s ability to acquire skills at a faster pace, brain-to-brain interaction for those with no underlying conditions and hands-free device control. GAO further adds that neural data is not covered by HIPAA outside of a clinical context because there isn’t a federal privacy law to fill in the gap. Most notably, the Department of Commerce is considering placing export controls on this technology - indication that there is massive potential beyond medical use cases.

BCIs read neural signals - so naturally it is prone to neural data privacy issues. This tech, unfortunately doesn’t have a basic password leak issue, instead it’s a company with access to someone’s movements, and plausibly their thoughts too. According to GAO, data collected for nonclinical purposes outside a clinical context isn’t covered by HIPAA.

Despite these, I see a lot of potential within BCI. Its still in its nascency but the tech could be applied in places we least think of - I’m thinking prisons or sports. Imagine the time saved interrogating suspects or better focus by knowing an athlete’s current state of mind. Of course, the application spans beyond these two examples and with the progress the startups are on will only amplify the use cases.