Illustration: Midjourney
When Vasileios Christopoulos, now an assistant professor in the Alfred E. Mann Department of Biomedical Engineering at USC Viterbi, was working in a brain-computer interface lab at Caltech, helping paralyzed patients control robotic arms with their thoughts, he kept hearing something he hadn’t expected. The patients weren’t asking about walking. They weren’t asking about their hands.
Dr. Charles Liu, professor of neurological surgery and neurology at the Keck School of Medicine of USC, had noticed the same thing across years of patient care. As the neurosurgeon who has implanted more brain-computer interfaces than anyone in the world, he’d heard it directly. “What became very obvious,” he says, “was that the patients wanted bladder function. It was actually more important.”
One patient stayed with Christopoulos. A woman who had written an entire book using only her mouth, before she ever had brain electrodes implanted.
“She told me, ‘Motor restoration is not a big deal for me,'” he recalls. “But I have a catheter on my bladder, because I cannot control it.” That catheter kept sending her to the hospital with kidney infections so severe the lab canceled experiments every other month. Because she couldn’t feel her bladder filling, she couldn’t feel herself getting sick, until the infection had reached her kidneys.
It was conversations like this that led to BLISS, (Bladder-Linked Stimulation System).
The problem nobody was solving
Approximately 308,000 people in the United States live with spinal cord injury. Nearly all lose bladder control. And yet the vast majority of research and engineering attention in neurotech has poured into motor restoration—making paralyzed limbs move again.
“When you can’t control your bladder, that’s all you think about,” says Liu, who also directs the USC Neurorestoration Center and is co-senior author on a new paper published in IEEE Transactions on Neural Systems and Rehabilitation Engineering. Lui explains odor is an issue. “It’s socially a huge problem. And medically, all of my brain-computer interface patients have a severe episode of urosepsis every year. I’ve known patients who died from this.”
Liu holds a PhD in bioengineering and trained under a pioneer who worked with Michael DeBakey on the artificial heart. He went to medical school, he says, for one reason: “I realized you can’t solve a problem you don’t understand. I went to medical school so I could be a better engineer.” That deliberate crossing of disciplines drove him to build a team around what he considers neuro-restoration’s most important unsolved problem.
A spinal cord machine interface
Brain-computer interfaces have made headlines for years. What this team is building is a spinal cord machine interface—and the distinction matters.
The spinal cord is a surprisingly elegant engineering target: specific fiber bundles carry specific signal types in consistent anatomical locations across individuals. And yet, Christopoulos notes, it has been largely dismissed by the neuroscience community. “When I was talking to people in neuroscience, most of the response was that it’s a cable.” Almost no functional neuroimaging research exists on the spinal cord—a striking gap for a structure that governs so much of human experience.
“The spinal cord is not just a cable,” says Shan Zhong, a postdoctoral researcher at USC Viterbi and first author on the paper. “Bladder control is sparsely distributed in the brain. But here, we can directly target one region and trigger the sense of bladder filling.”
The practical payoff matters too. Most existing technologies that help patients void rely on fixed schedules: an alarm goes off, and they catheterize. “The best thing about this,” Zhong says, “is that it can actually make people feel that there is a need for voiding, instead of depending on alarm clocks.”
That region is the dorsolateral funiculus, or DLF, a thin bundle of ascending sensory fibers near the spinal cord’s surface. Normally, as the bladder fills, signals travel up through the DLF to the brain, which registers fullness and sends a coordinated command back down: contract the bladder muscle and relax the sphincter simultaneously. After spinal cord injury, that loop is severed. The patient loses not just voluntary control but the ability to feel the need to go.
The team’s question: could you find the exact address of that signal, and replay it artificially?
Listening, then speaking
Using custom microelectrode arrays developed by Ecate LLC, a USC-affiliated startup, the team mapped neural activity in rats during controlled bladder filling. The arrays—with electrodes smaller than a human hair—were inserted into candidate spinal cord regions. Most were silent. But in the DLF, one or two adjacent channels lit up with rhythmic bursting that tracked filling precisely, climbing from 30 Hz with the first drops of saline to nearly 100 Hz just before voiding. Electrodes just 65 micrometers away stayed completely silent. The responsive zone, roughly 100 by 100 micrometers, was consistent enough across animals to serve as a reliable anatomical address.
Then the team spoke back. In a separate group of animals, they delivered patterned electrical pulses at those same coordinates, timed to mimic the biological signal of a full bladder. Coordinated voiding followed in 91.7% of trials, rising to 100% when the bladder was pre-filled to the volume where natural DLF activity begins. Leg muscle electrodes stayed silent throughout: the response was bladder-specific, not a generalized motor reflex.
The envisioned full system BLISS, for Bladder-Linked Stimulation System, would pair this sensory interface with a bladder volume sensor and a motor stimulator, creating a closed-loop neuroprosthesis that restores both the sensation and the act of voiding.
The pathway to patients
The team is already working with sheep, whose anatomy is closer to human scale. Liu estimates that with adequate funding, initial human recordings could begin within 18 months; not in spinal cord injury patients first, but piggybacked onto spinal cord tumor surgeries that are already far more invasive. A brief recording during an existing surgery adds minimal risk, while those patients often face bladder complications themselves and have a direct stake in what’s being built.
“Can we do it? Yes, the science is there,” Liu says. “Why are we doing it? Because the patients need it.” There is, he adds, “a very viable pathway” to a clinical trial. The first rat experiments were done on essentially no budget. “Alessandro Maggi (founder of Ecate) and I just said, ‘Let’s just do it.'” Federal grants, SBIR mechanisms, and an early-stage startup are all now in play. “We’re gonna get it done.”
The time of cross pollination
The paper lands alongside a significant institutional moment. On March 31, 2026, USC announced that the Alfred E. Mann Department of Biomedical Engineering will become a joint department between Keck and Viterbi, one of the first of its kind in California. BLISS is exactly what that partnership is designed to produce: a neurosurgeon who speaks engineering and a neural engineer shaped by years in clinical research, tackling a problem neither discipline could have framed alone.
“I’ve been at USC since 1996,” Liu says. “We have a great medical school and a great engineering school. But there didn’t seem to be a lot of cross-pollination. That’s a dream come true.”
“We want to restore the communication itself,” says Christopoulos. “Not just help them empty the bladder. We want them to feel normal again.”
“Intraspinal Microstimulation of Dorsolateral Funiculus for Coordinated Bladder Control” appears in IEEE Transactions on Neural Systems and Rehabilitation Engineering (DOI: 10.1109/TNSRE.2026.3675572). The work was supported by NSF grant 2403910 and the USC Neurorestoration Center.
Published on April 28th, 2026
Last updated on April 28th, 2026