On August 29, in the field of brain-machine interfaces (BMI), Elon Musk's Neuralink has been highly anticipated. However, a new small, thin-film chip from Switzerland has overshadowed Neuralink, offering not only a smaller size but also superior performance.

The chip developed by researchers at the École Polytechnique Fédérale de Lausanne (EPFL) represents a significant breakthrough in the field of brain-machine interfaces (BMI). BMI is a technology that can read brain activity and convert it into real-world outputs, such as text on a screen. The chip, named the Micromachine Brain Interface (MiBMI), is notably compact, consisting of two thin chips with a total area of just 8 square millimeters. In contrast, Elon Musk’s Neuralink device is relatively larger, measuring approximately 23 x 8 millimeters.

Additionally, the EPFL chip features extremely low power consumption, minimal invasiveness, and a fully integrated system capable of real-time data processing. This contrasts with Neuralink, which requires 64 electrodes to be implanted in the brain and processes data through an external device application.

Mahsa Shoaran, head of the EPFL Integrated Neural Technologies Laboratory, stated: “MiBMI enables us to convert complex brain activity into readable text with high precision and low power consumption. This advancement brings us closer to practical implantable solutions that could significantly improve communication abilities for individuals with severe motor impairments.”

Like other BMIs, MiBMI monitors electrical activity in the brain and converts it into outputs based on past brain monitoring data. MiBMI can read the brain signals formed when someone imagines drawing letters and convert these signals into text.

According to IT Home, the MiBMI chip has not yet been tested in live subjects, but it has achieved a 91% accuracy rate in converting neural activity into actual text, using real-time neural recordings collected from previous brain interface tests for training.

During the research, EPFL scientists identified a series of very specific neural markers that trigger when a patient imagines writing each letter. These markers are referred to as unique “neural codes” (DNC).

Currently, MiBMI can decode 31 different characters, which researchers claim is a new record for integrated systems of this kind. They believe that the system could eventually decode up to 100 different characters.

Researchers are also exploring other potential uses for the system, which may extend beyond text processing. Relevant studies have been published in the latest issue of the IEEE Solid-State Circuits Magazine.