AINEWS
Building Electronics the Brain Won't Reject
At Johns Hopkins, Xiao Yang is redesigning the brain implant from the cell up, part of a broader shift in neuroelectronics toward devices the body treats as native.
At Johns Hopkins, Xiao Yang is redesigning the brain implant from the cell up, part of a broader shift in neuroelectronics toward devices the body treats as native.
In a Johns Hopkins lab, a piece of electronics smaller than a single neuron sits next to a brain organoid in a dish. The device is soft. The neurons around it are softer. Xiao Yang leans in to read a measurement on a screen, watching whether the tissue treats the implant as part of itself or as something to wall off.
This is the practical problem at the center of Yang's work. Today's clinical brain implants, the rigid metal electrodes used to quiet the tremor of Parkinson's disease or to modulate circuits in treatment-resistant obsessive-compulsive disorder, are larger than the cells they sit beside. The body answers them the way it answers any foreign object: with scar tissue. That scar tissue is what eventually makes the implant fail.
Yang's group is trying a different approach. They build devices that are less than a micron across and shaped to mimic the structure and mechanical give of neurons themselves, according to a profile by Stephanie Pappas in Scientific American. The bet is straightforward. If an implant moves with the tissue and reads as something the tissue already recognizes, the brain will not attack it.
The problem Yang's lab is solving is older than Yang. Rigid electrodes became the default in clinical neurotechnology because they were the easiest to manufacture, not because they were the easiest for the body to accept. The scar response has been a known limitation of deep-brain stimulation hardware since the therapy became standard for movement disorders. What is new is the design choice being made by a small group of researchers, Yang among them: the implant's first job is to belong.
That choice carries a longer question Yang has framed in stronger terms. In the Scientific American profile, she describes a future of "a fully immersive and fully interactive hybrid system" between brain and machine. That vision language sits a long way from the bench work in front of her. The bench work is about making devices the body tolerates. The vision is about what a tolerated device is for.
That gap is the real story. Yang is not promising a merged future. She is solving an engineering problem that has to be solved before any such future is even discussable, and she is doing it by treating the implant as a kind of architecture inside the tissue, not as a probe stuck into it. Whether the larger hybrid vision is the right destination, or even a coherent one, is a question her generation of researchers is being asked to answer by building, and by choosing what to build.
The clinical stakes are concrete in the near term. Deep-brain stimulation already eases Parkinson's tremor and muscle rigidity. Electrode stimulation is used for treatment-resistant OCD. Both therapies work, and both are limited by the same thing: the body's response to a hard foreign object where soft native tissue should be. A device the body does not scar over could outlast the current generation of implants, deliver more stable recordings, and reach smaller targets. That part of the work can be measured.
The rest of it cannot. A sub-micron, neuron-shaped electronic is not a brain-computer interface in the consumer sense. It is a building block, and what it builds is still being decided. The boundary between device and tissue, the line Yang's generation is quietly redrawing, is not a technical fact. It is a design decision being made, piece by piece, in labs like hers.