Tiny sensors the size of a grain of sand could allow brain-machine interface control of prosthetics and implantable wearable tech.

"Neural dust" is a term used to describe tiny sensors designed by the EECS department of UC Berkeley. In a paper released this month, Berkeley researchers revealed that they’ve recorded the first in-vivo readings from implanted dust.

This research is a long time coming. In 2013, the team published research detailing their research on their use of ultrasound with CMOS circuitry. In 2015, they released another paper that further focused on theory, modeling, and scaling.

The resultant prototype in this most recent announcement is a step towards sensors that can be safely implanted in the brain. It's also a step towards a future where wearable technology could be implanted directly inside the body. 


The prototype neural dust device with a penny for scale. Screenshot courtesy of UC Berkeley.



The neural dust functions by using CMOS (complementary metal-oxide-semiconductor) technology. The CMOS component is needed to convert piezoelectric AC signals to DC via a full-wave bridge rectifier. In order to supply consistent and safe DC voltage to the CMOS, regulators are also necessary, along with DC-coupled ADCs and modulators.


A simplified version of the neural dust schematic. Image courtesy of the Cornell University Library.


Tiny, Batteryless Sensors

One of the largest challenges for any tiny sensor is power. In this case, the task was to provide power to CMOS circuitry small enough to measure in millimeters. In the case of neural dust, the prototype measures a mere 3 x 1 x 1 mm.


A neural dust "mote" on the tip of a finger. Image courtesy of UC Berkeley.


In addition to the issues associated with fabricating such small-scale circuitry, the neural dust has the massively important parameter of not producing appreciable heat while seated on a human brain.

The neural dust team dealt with the problem of power by utilizing ultrasound (PDF). Ultrasound waves emitted from outside the body are converted into electricity using a piezocrystal that feeds the resulting power to the transistor.


Graphical representation of the neural dust device. Image courtesy of UC Berkeley.


Ultrasound is also useful for this particular project because it’s capable of being transmitted and received pretty much anywhere in the human body. Where RF has limitations to how well it transmits within (and through) the human body, ultrasound is more robust. So not only does ultrasound solve the dust’s power issue, it also enables the communication of the device with monitoring equipment (the "interrogator") outside the body.


The directed ultrasound input (blue) and recorded backscatter (orange). Image courtesy of UC Berkeley.


Neural Control for Prosthetics

When the neural dust attaches to nerve fibers, it's capable of reading electrical impulses passing between neurons via electrodes. The ability to measure these impulses is crucial in developing an electromechanical system that can respond to them and physically move a prosthetic.


The neural dust prototype attached to a nerve fiber in a rat. Image courtesy of UC Berkeley.


The goal is to have neural dust feed impulses to a receiver that in turn moves the mechanical portion of a prosthetic. This would allow amputees to be able to control a replacement limb merely by thinking about it.

Implanted "Wearables"

There are myriad applications these sensors could have in the medical field beyond control of prosthetics. MEMS (microelectromechanical systems) technology is a popular subject of research at present. Recently, scientists from multiple fields have developed projects like prototype brain implants.

Of course, there are plenty of commercial applications the neural dust could enable. Sometime in the future, we could see a generation of wearables that are implanted directly inside the body. This could allow real-time data on organ health, giving insight into systemic health and even advance warning for heart attacks, strokes, and other emergencies.

But along with the utility of wearables also comes the danger of security breaches. To that end, you may have heard of smart dust in the past, though possibly in a much more sensational context. In 2013, MIT’s Technology Review published an article titled “How Smart Dust Could Spy On Your Brain”. It points out that the same features that make neural dust appealing for medical purposes (mobile tracking, remote monitoring, etc.) also make it appealing for data collection.

Therefore, unfortunately, the same security issues that plague wearable technologies, both against hackers and marketers, are likely to remain when wearables become "implantables". These concerns are likely a ways off, however, since the team at UC Berkeley are still in the process of developing the neural dust design.



The neural dust program is helmed by the EECS program at Berkeley and funded in part by DARPA.

Learn more about the neural dust program here.