Communications of the ACM
Ultra-Low-Power Brain Implants Find Meaningful Signal in Gray Matter Noise
Researchers at the University of Michigan have designed an ultra-low-power brain implant with enhanced accuracy.
Credit: University of Michigan News
University of Michigan (UMich) researchers have reduced the power requirements of neural interfaces by an estimated 90% and enhanced their accuracy in identifying intention by designing an ultra-low-power brain implant.
The team compressed brain signals by exclusively monitoring spiking-band power (SBP), an integrated set of frequencies from multiple neurons ranging from 300 Hz to 1,000 Hz. This facilitated highly accurate prediction of behavior with dramatically reduced power needs.
The method is as accurate as transcutaneous systems while requiring only a tenth as many signals, and can predict neuron firing more accurately, even amid gray matter noise.
UMich's Samuel Nason said this achievement opens up many existing devices to brain-machine interface applications.
From University of Michigan News
View Full Article
Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA
No entries found