Developing high data rate and low-power neurotechnology using backscatter communication

Aleenah Ansari

Researchers in Matt Reynolds’ lab at the University of Washington are developing implantable brain-computer interfaces that have increased capacity for transferring power and data wirelessly.

If you’ve ever scanned a badge, transit pass or entry card, you’ve probably used a well-established communications technology called radio frequency identification (RFID), which transfers data using minimal energy. Center for Neurotechnology (CNT) member, Matt Reynolds, and researchers in his CNT-affiliated lab at the University of Washington (UW) are using both RFID and backscatter communication, a relatively new method of wireless communication with high-data transfer and low-power consumption, all without batteries and wires. Backscatter communication can be used in everything from internet-enabled appliances to security.

In close partnership with CNT member, Joshua Smith, and his Sensor Systems Laboratory at the UW, Reynolds and his research team are using backscatter communication to develop implantable brain-computer interfaces (BCIs) that can be used in future research and medical applications.

“Researchers want as much data as possible, but they also want long durations [of data collection [to understand the dynamics of neuroplasticity over time],” said James Rosenthal, a CNT and Reynolds’ lab member and third-year PhD student in the UW Department of Electrical & Computer Engineering. “These goals require compromise. More data requires more power. How can we efficiently give [researchers] as much data as possible?”

To understand how backscatter communication works, imagine being in a dark room where you have a flashlight, and somebody else has a mirror. Your goal is to communicate with each other, and reflecting or not reflecting the light can be a signal. From an electronics perspective, it would take a significant amount of power to run the flashlight, but the mirror requires almost no power.

“Backscatter communication doesn’t actively transmit information. It reflects data, so you can reduce the power consumption of the wireless link by hundreds of times,” said Lefteris Kampianakis, CNT member, researcher in Reynolds’ lab and PhD student in the UW Department of Electrical & Computer Engineering.

This research on backscatter communication aligns with the CNT’s goal of developing innovative neural devices that engineer neuroplasticity, a new form of rehabilitation that uses engineered devices to restore lost or injured connections in the brain, spinal cord and nervous system. BCIs are first being used by researchers, and as the technology improves over time, will begin to make their way into clinical applications for people impacted by spinal cord injury, stroke and other neurological disorders.

Designing with constraints on size, weight and power

Currently, Rosenthal, Kampianakis and other researchers in Reynolds’ lab are developing a wireless device that can send samples of neural data from a freely-moving non-human primate. To do this, Rosenthal uses novel backscatter designs to increase data rates. Thanks to lower power consumption, the backscatter communication system ensures that researchers can transmit and record higher data rates and extend experiment duration. Extending the battery life also reduces the frequency of surgeries to replace an implanted device, which is beneficial for patients.

“In the end, you don’t want a patient to have a huge battery,” Kampianakis said. “It’s the same reason you want your cell phone to last all day.”

Rosenthal is also focused on providing scientists with data at a low-frequency perspective, which shows how neuronal populations change over time, and a high-frequency perspective, which shows how individual neurons behave.

“Both populations are of interest,” Rosenthal said. “We want to provide both so researchers can look at both spatial and temporal resolutions.”

Rosenthal worked at the NASA Langley Research Center from 2013 to 2018. There, he had to consider constraints on size, weight, and power when developing electronic systems for science and technology development missions, which is also true of his current research.

“It’s expensive to send big things up into space. You’re very limited in what you have. With neural recording, it’s the same thing,” Rosenthal said. “Since you don’t want to impede natural behavior, devices have to be very small, which means the batteries are small too.”

For this research, Apoorva Sharma, a CNT member, researcher in Reynolds’ lab, and fourth-year PhD student in the UW Department of Electrical & Computer Engineering, focuses on creating dual-band antennas for BCIs that support wireless power transfer and communication.

“A conventional BCI is connected to the computer with wires for power and for communication,” Sharma said. “This works for experiments, but it’s not good for everyday use. If we can transmit data wirelessly, we can use this system for daily life.”

To design the antenna, Sharma had to understand the material properties of non-human primate tissue and how it transmits electric force without conduction. She focused on designing antennas to reduce absorption rates of electromagnetic radiation, avoid tissue heating, and decrease the amount of interference between power and communication antennas.

Using interdisciplinary collaboration to reduce signal interference

The CNT has invested in making sure that the research in Reynolds’ lab is a collaborative effort by bringing in stakeholders from electrical engineering, computer science, bioengineering and philosophy. This research project required five different programming languages to ensure that the wireless power transfer, antenna design, and MATLAB receivers all worked together.

One of the recent milestones of this research was the successful measurement and uplink of broadband neural data at 25 megabits per second while only consuming 12.4 picojoules per bit, which was four times more neural data than the research team’s previous work.

Reynolds‘ lab works with the Washington National Primate Research Center and follows their ethical protocol for implanting BCIs in primates, which is a stepping stone to developing BCIs for humans. The researchers are cognizant of the challenges involved in collecting data from a BCI that is implanted in a freely- moving primate.

“Primates put their hands on the antenna and move around, so you have to design the system to be robust when they move their head in different directions,” Rosenthal said.

Another challenge is the fact that the primates are in metal cages, so signals can bounce off the walls and interfere with signal transfer from the BCI. To address this, researchers in the lab created an adaptive cancellation system that responds to this interference. To develop this system, Kampianakis used adaptive circuitry so the device would respond to environmental changes in the cage, and Sharma is working on understanding electromagnetic changes with the goal of reducing the amount of interference.

Long-term, the research team hopes to create implantable BCIs that can use backscatter communication and have longer battery life. This could translate to fewer surgeries for patients who use BCIs as a treatment for neurological conditions such as spinal cord injury, essential tremor and Parkinson’s disease.

“It’s exciting to contribute to research in engineered neuroplasticity that could transform how we treat brain injuries and [neurological] disorders,” Rosenthal said.

For more information about this CNT-funded research, contact James Rosenthal.