Developing StiMote, a wireless neural stimulator for restoring vision

The highly collaborative project will leverage many tiny sensing computers, called “motes,” to communicate with the visual cortex of the brain.

An interdisciplinary team of experts across the University is developing wireless sub-mm-scale devices for stimulating neurons, with the goal of restoring vision. The team, which includes Electrical and Computer Engineering (ECE) faculty David Blaauw and Biomedical Engineering faculty Cindy Chestek and James Weiland, call the stimulation devices StiMotes, a nod to the world’s smallest computer, known as the Michigan Micro Mote (M3).

“We’re making tiny little independent motes that float right on top of the tissue of the brain,” explained Blaauw, the Kensall D. Wise Collegiate Professor of Electrical Engineering and Computer Science. “They are so small, the brain doesn’t even notice that anything is there.”

A quartz-like rock, which is really a zoomed-in grain of sugar, sits next to a tiny black chiplet with a wire extending from it. The scale of the chiplet is 150 micrometers.
Early prototype of neural recording mote next to a grain of sugar. Image courtesy of David Blaauw.

The current technology used to restore vision stimulates retinal cells at the back of the eye, allowing blind patients to see patterns of light and dark. With this stimulation, patients can learn to distinguish and interpret familiar edges or shapes in the environment, like doorways or furniture. Because the area of the retina is so small and the cells are so densely packed, there is a limit to the resolution that can be achieved. Retinal stimulation is also ineffective for patients with damage to the optic nerve.

The StiMote system is an alternative to retinal stimulation—it will stimulate the visual cortex at the back of the brain, where the neurons that control vision are distributed across a larger area. To stimulate such a large area, researchers would traditionally use a rigid implant with many electrodes; however, these implants are unable to adjust to any tiny movements of the tissue and ultimately cause damage and scarring to the cells. Instead, each StiMote is implanted with a carbon fiber wire over 10 times smaller than an electrode, enabling the team to electrically stimulate the neurons without damaging the tissue.

A grid of four images. The top two, in black and white, show the damage that the Utah array leaves as an approximately 100 micrometer hole in the brain. The bottom row, in teal, highlights the lack of damage from the carbon fiber wires used by the StiMote team.
Top: A hole left in brain tissue from the implantation of a Utah Array electrode to stimulate brain cells (L), next to a control area of the brain without implantation [source]. Bottom: Six carbon fiber wires implanted across the midline of a brain section (L), next to a control area of the brain without implantation [source]. The carbon fiber wires do not damage the tissue in the same way as the Utah Array electrodes. Image adaptation courtesy of Cindy Chestek

“We can assemble the StiMotes into arrays in a custom fashion. Everyone’s brain is different in subtle ways and so being able to customize the array for each person’s brain is going to lead to a better outcome and allow us to specifically target an area in a much more precise way,” said Weiland. “I’m excited about the prospect of artificial vision, but also that some of the technology we’re using has a broad application throughout neuromodulation. Any place on the brain that you would want to stimulate or record from can benefit from the mote approach.”

Because the StiMote system is wireless, the team is removing the most likely failure mechanism for traditional brain-machine interface implant surgeries.

“The hardest part of the existing surgery to implant a device is managing the bundle of wires that comes off the electrode,” said Chestek, “Having those wires tethering the electrodes to some distant medical device is the primary failure mechanism for the current technologies, and the hardest part to keep working for 10 years, or however long the implant needs to last. So we’re getting rid of that altogether.”

Chestek originally enlisted Blaauw’s assistance to build a device capable of recording the activity of the cells; that device was called ReMote (i.e., recording mote). After becoming interested in stimulation, they contacted Weiland to contribute his expertise to the StiMote project.

“The problem with stimulation is that you need to pump quite a bit of energy into the brain to stimulate a neuron response,” said Blaauw. “Because of technology improving, we’re able to put a very small capacitor on there, about the width of a hair. It can store quite a bit of charge, and it can store enough charge that it can stimulate the neuron to fire once. We can charge them very slowly and have that charge available all at once when we need to actually stimulate.”

To charge the capacitors and communicate with the individual motes, the system uses a repeater that sends infrared light signals. Infrared light is safe for the brain tissue, but, Blaauw said, “chips generally don’t like light.”

To overcome this obstacle, Blaauw altered the design of the entire chip to be light-tolerant. 

“A philosophy we’ve tried to use is that we need to make the circuits subservient to the system; we need to make circuits that make the whole system feasible, even when it creates unique circuit challenges,” he said.

To stimulate the visual cortex and create artificial vision, the team will implant hundreds or thousands of individual StiMotes to float just above the surface of the brain, like balloons tethered by their fiber optic wires.

A diagram of the brain with a grid of stimulation motes, next to a zoomed in circuitry diagram of a single mote and a diagram of the mote dimensions.
A conceptual overview of the proposed free-floating stimulator motes and cross-sectional view of stacked layers. Image from the paper, “A Wireless Neural Stimulator IC for Cortical Visual Prosthesis.”

StiMote has the potential to revolutionize brain stimulation in patients, restoring vision loss, simulating feeling in prosthetic limbs, and enabling speech for people with paralysis.

The team is far from achieving this vision, but, Chestek said, “This is a better version of the technology than they’re currently using. If this project works, it will be a giant step forward.”

Blaauw, Chestek, and Weiland have ongoing meetings with about 20 researchers across the university, and as far away as ETH Zurich. They include surgeons, chip designers, and industrial contacts. An earlier version of the technology was described in the 2023 paper, “A Wireless Neural Stimulator IC for Cortical Visual Prosthesis.” 

ECE Professors Hun-Seok Kim and Dennis Sylvester, as well as former ECE Professor Jamie Phillips, helped develop the hardware for the StiMotes. 

In addition to their core appointments in Biomedical Engineering, Weiland is a professor of Ophthalmology and Visual Sciences, and Chestek is a professor of Robotics and ECE.