Miniaturized brain-machine interface processes neural signals in real time

EPFL researchers have developed a next-generation miniaturized brain-machine interface capable of communicating directly between the brain and text on tiny silicon chips.

Brain-machine interfaces (BMIs) have emerged as a promising solution to restore communication and control in people with severe motor disorders. Traditionally, these systems have been bulky, power-hungry, and limited in their practical applications.

EPFL researchers have developed the first high-performance miniaturized brain-machine interface (MiBMI), offering an extremely small, low-power, highly accurate and versatile solution.

Published in the latest issue of IEEE Semiconductor Circuits Journal and presented at the International Solid-State Circuits Conference, MiBMI not only improves the efficiency and scalability of brain-machine interfaces, but also paves the way for practical, fully implantable devices.

This technology has the potential to significantly improve the quality of life of patients with diseases such as amyotrophic lateral sclerosis (ALS) and spinal cord injuries.

The small size and low power consumption of the MiBMI are key features that make it suitable for implantable applications. Its minimally invasive nature ensures its safety and practicality for use in clinical settings and in real life.

It is also a fully integrated system, meaning that recording and processing are done on two extremely small chips with a total area of ​​8 mm.²It is the latest in a new class of low-power BMI devices developed at Mahsa Shoaran’s Integrated Neurotechnologies Laboratory (INL) at EPFL’s IEM and Neuro X institutes.

“MiBMI allows us to convert complex neural activity into readable text with high accuracy and low power consumption. This advance brings us closer to practical, implantable solutions that can significantly improve the communication abilities of people with severe motor disorders,” Shoaran says.

Brain-to-text conversion involves decoding the neural signals generated when a person imagines writing letters or words. In this process, electrodes implanted in the brain record neural activity associated with the motor actions of handwriting.

The MiBMI processor then processes these signals in real time, translating the brain’s predicted hand movements into corresponding digital text. This technology allows individuals, particularly those with locked-in syndrome and other severe motor disorders, to communicate simply by thinking about writing, with the interface converting their thoughts into readable text on a screen.

“Although the chip has not yet been integrated into a working BMI, it has processed data from previous live recordings, such as those from the Shenoy lab at Stanford, converting handwriting activity to text with an impressive 91% accuracy,” says lead author Mohammed Ali Shaeri.

The chip can currently decode up to 31 different characters, a feat unmatched by any other embedded system. “We are confident that we can decode up to 100 characters, but a handwriting dataset with more characters is not yet available,” Shaeri adds.

Current BMIs record data from electrodes implanted in the brain, then send these signals to a separate computer for decoding. The MiBMI chip not only records the data but also processes the information in real time, integrating a 192-channel neural recording system with a 512-channel neural decoder.

This neurotechnological breakthrough is a feat of extreme miniaturization that combines expertise in integrated circuits, neural engineering and artificial intelligence. This innovation is particularly interesting in the emerging era of neurotechnological startups in the field of neural integration, where integration and miniaturization are key axes. EPFL’s MiBMI offers promising perspectives and potential for the future of the field.

To be able to process the enormous amount of information captured by the electrodes of the miniaturized BMI, the researchers had to adopt a completely different approach to data analysis. They discovered that the brain activity of each letter, when the patient imagines writing it by hand, contains very specific markers, which the researchers called distinctive neural codes (DNC).

Instead of processing thousands of bytes of data for each letter, the microchip only needs to process the DNCs, which are about a hundred bytes. This makes the system fast and accurate, with low power consumption. This advancement also allows for faster training times, making learning to use the BMI easier and more accessible.

Collaborations with other teams at EPFL’s Neuro-X and IEM institutes, including the labs of Grégoire Courtine, Silvestro Micera, Stéphanie Lacour, and David Atienza, promise to create the next generation of integrated BMI systems. Shoaran, Shaeri, and their team are exploring various applications of the MiBMI system beyond handwriting recognition.

“We are collaborating with other research groups to test the system in different contexts, such as speech decoding and movement control. Our goal is to develop a versatile BMI that can be adapted to various neurological disorders, thus offering a wider range of solutions to patients,” Shoaran explains.