Swiss Researchers Develop Tiny Brain-Computer Interface with 91% Accuracy in ConvertingThoughts to Text
Zurich, Switzerland – A team of researchersat the École Polytechnique Fédérale de Lausanne (EPFL) in Switzerland has made a significant breakthrough in the field of brain-computer interfaces (BCIs) with the development of a miniature, high-performance chip called the Micro Brain-Computer Interface (MiBMI). This tiny device, measuring just8 square millimeters, outperforms Elon Musk’s Neuralink in both size and accuracy, demonstrating the potential to revolutionize communication for individuals with severe motor impairments.
BCIs are devices that can read brain activity and translate it into real-world outputs, such as text on a screen. The MiBMI, consisting of two thin, flexible chips, is significantly smaller than Neuralink’s device, which measures approximately 23 x 8 millimeters. This smaller size translates toless invasiveness and a lower risk of complications during implantation.
Furthermore, the EPFL chip boasts a low power consumption and a fully integrated system for real-time data processing. This contrasts with Neuralink, which requires 64 electrodes to be inserted into the brain and relies on external devices and applications fordata processing.
MiBMI allows us to translate complex brain activity into readable text with high accuracy and low power consumption, explained Mahsa Shoaran, head of the EPFL’s Integrated Neurotechnologies Laboratory. This advancement brings us closer to practical implantable solutions that can significantly improve communication abilities for individuals with severemotor disabilities.
The MiBMI, like other BCIs, monitors electrical activity in the brain and converts it into output based on previously collected brain monitoring data. The device can read the brain signals generated when a person imagines writing letters and output those signals as text.
While the MiBMI chip has not yet beentested in living subjects, it has been trained using real-time neural recordings collected from previous brain interface tests, achieving a remarkable 91% accuracy in converting neural activity into actual text.
During their research, the EPFL team discovered a series of highly specific neural markers that are triggered when patients imagine writing each letter. They refer to these markers as unique neural codes (DNCs). Currently, the MiBMI can decode 31 different characters, a new record for similar integrated systems. The researchers believe they can eventually expand the system to decode up to 100 different characters.
The team is now exploringother potential applications for the MiBMI, which could extend beyond text processing. The research has been published in the latest issue of the IEEE Journal of Solid-State Circuits.
This groundbreaking development holds immense promise for individuals with conditions such as amyotrophic lateral sclerosis (ALS), stroke, or spinal cord injuries, who struggle tocommunicate effectively. The MiBMI’s ability to translate thoughts into text could empower these individuals to regain their voice and connect with the world around them.
The research also highlights the rapid advancements in the field of brain-computer interfaces, pushing the boundaries of what is possible in human-machine interaction. As these technologies continueto evolve, they have the potential to transform our understanding of the brain and revolutionize how we communicate and interact with the world.
【source】https://www.ithome.com/0/791/964.htm
Views: 0