Salt-Sized Sensors Mimic the Brain

To achieve a greater understanding of the mind, why not draw inspiration from it? No less than, that’s what researchers at Brown College did, by constructing a wi-fi communications system that mimics the mind utilizing an array of tiny silicon sensors, every the dimensions of a grain of sand. The researchers hope that the know-how might sooner or later be utilized in implantable brain-machine interfaces to learn mind exercise.

Every sensor, measuring 300 by 300 micrometers, acts as a wi-fi node in a big array, analogous to neurons within the mind. When a node senses an occasion, comparable to a change in temperature or neural exercise, the system sends the info as a “spike” sign, consisting of a sequence of quick radiofrequency pulses, to a central receiver. That receiver then decodes the data.

“The mind is exquisitely environment friendly in dealing with giant quantities of information,” says Arto Nurmikko, a professor of engineering and physics at Brown College. That’s why his lab selected to develop a community of unobtrusive microsensors which are “neuromorphic,” that means they’re impressed by how the mind works. And the similarities don’t finish there—Nurmikko says that the wi-fi alerts and computing strategies are additionally impressed by the mind. The group published their results on 19 March in Nature Electronics.

Considering Like a Mind

Like neurons, these sensors are event-driven and solely ship alerts to the receiver when a change happens. Whereas digital communication encodes info in a sequence of ones and zeros, this method cuts down the quantity of information transmitted by utilizing intervals of inactivity to deduce the place zeros can be despatched. Importantly, this results in vital power financial savings, which in flip permits for a bigger assortment of microsensors.

However with so many sensors sending info to a typical receiver, it may be troublesome to maintain the info streams straight. The researchers deployed a neuromorphic computing approach to decode the alerts in actual time.

“The mind is exquisitely environment friendly in dealing with giant quantities of information.” —Arto Nurmikko, Brown College

The researchers additionally performed simulations to check the system’s error fee, which will increase with extra sensors. Along with 78 fabricated sensors, they ran simulations of networks consisting of 200, 500, and 1,000 nodes utilizing an actual knowledge set from primate mind recordings. In every, the system predicted the hand motion of a non-human primate with an error fee beneath 0.1 %, which is appropriate for brain-computer purposes. Nurmikko says the group will subsequent check the wi-fi implanted sensor community in rodents.

Whereas the know-how could possibly be utilized to any a part of the physique the place biomedical researchers intention to observe physiological exercise, the first purpose is use in a brain-machine interface that may probe a big area of the mind, says Nurmikko. The sensors is also modified to be used in wearable technology or environmental sensors.

There are key benefits of the system for biomedical makes use of, such because the small, unobtrusive design. However these purposes additionally impose a key limitation: The sensors are externally powered by a wireless beam to keep away from the necessity for batteries, and the physique can only safely absorb so much radiofrequency energy. In different phrases, the system will not be restricted by bandwidth, however as an alternative by energy supply. “From a sensible viewpoint, it all the time comes again to the query of, the place do you get your power?” says Nurmikko.

Mind-Machine Interface Prospects

The analysis supplies “an vital contribution, which demonstrates the feasibility and potential of neuromorphic communications for future use instances of low-power wi-fi sensing, communication, and resolution making,” says Osvaldo Simeone, a professor at King’s School London and one of many researchers who first designed and simulated a neuromorphic communication system in 2020.

The concept of a wi-fi community probing the mind will not be new, says Federico Corradi, a researcher and assistant professor {of electrical} engineering at Eindhoven College of Expertise. In 2011, for instance, a researcher at UC Berkeley gave a presentation on “neural dust” during which he proposed a hypothetical class of nanometer-sized wi-fi sensors. “However now, it’s materializing slowly,” Corradi says.

One vital factor of the Brown researcher’s design is its simplicity, says Corradi. The sensor’s structure doesn’t embrace a battery or clock embedded throughout the chips, making it splendid for scalable, low-power methods. “It opens quite a lot of prospects.”

Moreover, Corradi factors to the sensor’s asynchronous nature as a key benefit—and limitation. This side of the sensor preserves time info, which is crucial for learning the mind. However this function might additionally introduce issues if the relative timing of occasions will get out of whack.

Corradi believes this work is a component of a bigger pattern towards neuromorphic methods, a “new wave of brain-machine interfaces that I hope we are going to see within the coming future.”