Scientists Grew Soft, Wire-Free Electrodes Inside Living Brains — Using Only a Drop of Blood
The mice are fine. That is the first thing to say about a Purdue University experiment that sounds like it should go wrong in several ways at once.
In a procedure that would make a neurosurgeon wince, researchers injected a liquid monomer into the motor cortex of living mice, then waited. What formed over the next few hours was not a scar or an immune reaction. It was an electrode: a soft, flexible mesh of conductive polymer that wrapped around individual neurons. The catalyst was not a toxic chemical or an implanted metal wire. It was the iron in hemoglobin — the same molecule that carries oxygen in blood.
The work appears in Science (April 2, 2026). Its most striking implication is practical: electrodes that grow in place, require no hardware to be inserted, and can be controlled from outside the skull using near-infrared light. No surgery. No chronic foreign-body response. The body builds its own interface.
"That is the future — growing soft, wire-free electronic interfaces inside the brain using the patient's own blood," Krishna Jayant, the Leslie A. Geddes Assistant Professor at Purdue's Weldon School of Biomedical Engineering and one of the corresponding authors, said in a university news release.
The monomer is a compound called BDF. When mixed with a drop of blood and delivered into brain tissue, the iron in heme proteins — hemoglobin in red blood cells and myoglobin in muscle — catalyzes a polymerization reaction that chains the monomers into a mesh. The mesh is electrically conductive and soft enough to move with the brain rather than against it. Previous approaches to in-brain electrode formation required copper-based catalysts, which are toxic to biological tissue. "Biologic systems don't like copper," Helen Tran, a bioelectronics researcher at the University of Toronto who was not involved in the study, told Chemistry and Engineering News. "However, the body has iron readily available in blood proteins like hemoglobin." Her concern is what happens next: "It takes time for the polymers to form and over time the polymer may degrade; we need to make sure these periods of time do not elicit unpredictable and adverse biological responses."
The mesh responds to light. Near-infrared pulses delivered from outside the skull selectively heat the polymer, triggering a mechanism the team calls thermionic modulation — acting directly on ion channels in individual neurons rather than on whole membranes. This allows millisecond-precision control of neural firing, the kind of temporal resolution needed to track and modulate the oscillatory activity associated with movement disorders and seizures. The existing alternatives — deep brain stimulation implants, for example — use implanted electrodes that mechanically mismatch brain tissue and generate chronic inflammation.
Over 80 percent of zebrafish embryos survived after in-vivo electrode formation in a parallel set of experiments and developed normally, swimming actively. Zebrafish are a common proxy for developmental neurotoxicity screening because their embryos are transparent and fast-developing.
The applications the team highlights are conditions marked by excessive or runaway brain activity: epilepsy, Parkinson's disease, chronic pain, and some forms of depression and addiction. These are conditions where existing neuromodulation devices have limited reach precisely because implanted hardware causes tissue damage over time. A grown-in-place mesh that moves with the brain could, in theory, maintain a stable interface for years without the drift and inflammation that plague rigid implants.
Helen Tran's caution about degradation is the right one to hold onto here. The electrode formation takes hours. Whether the mesh remains stable and functional over months or years inside a human brain is not known. Mouse brains are not human brains. The safety and toxicity data — what the degradation products are, whether they accumulate, what the long-term immune response looks like — will take a dedicated multi-year study to establish. This is early work.
What makes it worth watching is the conceptual shift. Existing brain-computer interfaces treat biology as the thing you have to work around: insert hardware, manage inflammation, accept degradation. This approach asks whether biology can be recruited as part of the interface itself. The iron in hemoglobin is already there. The polymerization reaction is already possible. What the Purdue team showed is that the two can meet in the brain without destroying each other. Whether that changes what a neural interface looks like in 20 years depends on data that does not yet exist.
The lead authors are Sanket Samal and Shulan Xiao, postdoctoral scientists at Purdue. Jianguo Mei, Richard and Judith Wien Professor of Chemistry at Purdue, is the other corresponding author alongside Jayant.
