In late May 2026, the website of Science magazine published an in-depth feature on Bexorg, a US biotechnology startup. The disclosed information turned the company — until then a low-profile spinoff out of Yale University — almost overnight into one of the most contested entities at the intersection of bioethics and neuroscience.
Founded in 2021 in New Haven, Connecticut, Bexorg was co-founded by Zvonimir Vršelj, a physician-scientist from the Yale School of Medicine who serves as CEO, and the neuroscientist Nenad Šestan. They have developed a proprietary ex-vivo brain perfusion device capable of sustaining the cellular metabolic activity of an intact human brain for roughly 24 hours after a donor's death, by pumping an oxygenated synthetic blood substitute through it. By the time of the Science report, Bexorg had already processed more than 700 intact human brains from deceased donors, for use in drug testing and neuroscience research.
Bexorg's technical framework makes no value judgment of its own — but it has dragged several subjects that used to belong to separate domains onto a single plane all at once: the experimental limits of neuroscience, the boundary-testing of bioethics, and the trust on which the organ-donation system rests. More than 700 intact human brains kept alive outside the body, together with the 1.2-meter-tall slicing robot about to be installed, compose a picture that leaves the scientific community at once thrilled and uneasy.
Technical Principles and Scale
Bexorg's core technical breakthrough lies in finding a way to keep an intact human brain from a deceased donor metabolically active outside the body. This is not "preservation" but "perfusion" — continuously delivering oxygen and nutrients to the brain tissue while clearing away metabolic waste, by mimicking the body's own circulatory system.
Each brain has a survival window of roughly 24 hours after perfusion begins. Within that window, researchers can test the direct effects of candidate drugs on intact brain tissue — changes in neuronal signaling, synaptic activity, gene-expression profiles — data that cannot be obtained at the same level of physiological fidelity from conventional cell models or animal models.
The company's laboratory is equipped with a 1.2-meter-tall robotic arm that automatically performs the slicing of up to 1,600 brain tissues a year and analyzes 11,000 proteins in each slice. Bexorg's commercial selling point is this: the ability to test therapeutic drugs directly on physiologically intact human brains, yielding translational data that animal models cannot provide.
The Source of 700 Brains and the Ethical Architecture
The company states publicly that all of its brains come with authorization from organ donors or their families, and that none are obtained involuntarily. Brain tissue is removed within hours of a donor's death. The company maintains a dedicated ethics advisory committee.
Yet the number 700 itself raises a practical question: even if every individual case has obtained lawful informed consent, once the volume processed reaches the hundreds — and potentially thousands in the future — whether "consent" as an ethical act can still retain its ethical force at the operational level remains an open question. The Science report notes that during media events Bexorg displays its assembly-line process while simultaneously trying to reassure the public that these brains, detached from the body, neither cross ethical lines nor carry any risk of restored consciousness. The company needs this two-way communication to stabilize outside expectations — demonstrating technical capability to attract investment and market interest, while repeatedly affirming its ethical boundaries to guard against a backlash in public opinion.
In traditional medical ethics, the handling of a deceased person's remains is bound by strict cultural, religious, and legal constraints. The brain's special status lies in the fact that it is the conceptual anchor of the "self" — unlike the heart or the liver, it is the physical substrate from which consciousness arises. When an intact human brain continues to sustain metabolic activity outside the body after death, both the definitional boundary of "death" and the ethical status of the remains become blurred.
Scientific Value and Commercial Pathway
From a pharmaceutical company's perspective, the service Bexorg offers fills a critical gap. One conspicuous problem in drug development today is this: in-vitro experiments (cell cultures) cannot accurately predict how a drug will perform inside the human body, while the species differences between animal models and humans frequently distort the data from preclinical research. Bexorg's human-brain perfusion system provides an intermediate model between the two — testing a drug's neural effects directly in intact human brain tissue.
This model could prove especially significant for the development of drugs targeting neurological diseases — Alzheimer's, Parkinson's, brain tumors, and the like. Pharmaceutical giants spend tens of billions of dollars in these areas every year, yet the translational success rate of new therapies remains stubbornly low. If a human-brain perfusion model could screen candidate drugs more accurately, its commercial value would far exceed that of conventional in-vitro screening platforms.
Where the Boundaries Lie
An ex-vivo perfusion system handling 700 human brains has placed several fundamental questions back on the table.
- An upgrade of the experimental subject. Moving the experimental subject from mouse brains and human-derived organoids to intact human brains from deceased donors is a leap with ample scientific justification — but ethics has no ready-made framework for discussing it. A human brain is not simply one more tissue that can be filed under the category of "biological sample."
- The unresolved question of "risk of restored consciousness." This is the very question Bexorg has been proactive in rebutting, and in fact it points to the sharpest theoretical dispute at the frontier of neuroscience: if an ex-vivo, living human brain approaches the state of a living body in its neuronal activity, how can researchers be certain that it has no conscious experience? The company currently claims to have ruled out this risk by technical means, but verification is difficult on both the technical and the theoretical level.
- The ethical pace of commercialization. As the technology expands from research use to commercial use, can ethical review keep pace with the tempo of scaled-up operation? Bexorg's capacity to slice 1,600 brain tissues a year poses a question of synchronization in pace.
Bexorg's scenario bears a structural resemblance to the problem of "socio-technical disjunction" in the field of AI safety ethics: the speed of technological development far outpaces the update cycle of ethical and regulatory frameworks. The "technology first, ethics catches up" pattern has been debated in the AI field for several years; in biotechnology, Bexorg has used the scale of 700 brains to turn it from a theoretical proposition into a concrete, present-day problem.