On October 26, 2026, MIT announced LUCID, an implantable device that injects substances directly into the brain and uses ultrasound to help them distribute into the surrounding tissue. It combines a thin injection tube with 24 ultrasound transducers. MIT reports that in tissue-mimicking gel and in mice, substances spread farther than they did without ultrasound. The work pairs a route for delivering substances to the brain with a mechanism that helps them spread after delivery.
The research was led by MIT's Canan Dagdeviren, with Sarah Ornellas and Jason Hou as lead authors. MIT announced it as research based on the original paper "Long-term ultrasound-enhanced convective injection device for neural drug delivery," published in the peer-reviewed journal Device. The bibliographic entry on BrightSurf lists the publication date as the same day and the DOI as 10.1016/j.device.2026.101333. However, these results compare how substances were distributed. They do not report that brain tumor recurrence was prevented.
After delivery to the brain, how do you spread it around?
Drug delivery to the brain is limited by the blood-brain barrier, which restricts the passage of substances from the blood into brain tissue. Injecting directly into the brain places a substance without going through that barrier. But the challenge remains of getting the substance from the injection site across the needed area.
In MIT's announcement, Hou said that getting a drug into the brain is only part of the challenge. According to his explanation, the team combined a delivery channel with a small ultrasound array, aiming to spread the substance into the surrounding tissue even during slow infusion.
In LUCID, a cannula, the thin tube that injects liquid, is 3D-printed directly onto an array of 24 small ultrasound transducers. The device is about as thin as cardstock and is flexible. Ultrasound is generated near the tip of the tube to help move the substance after it is released. MIT explains, citing prior findings, that ultrasound creates "convection," in which flow carries the substance and promotes its distribution.
The device's reservoir holds about 100 microliters of liquid, and an osmotic pump supplies it. What is wirelessly controlled over Bluetooth is the ultrasound, and the small battery can be charged wirelessly. The pump that supplies the liquid and the ultrasound that aids distribution have different roles. It has not been announced that injection volume or rate can be freely changed over Bluetooth.
MIT also mentions ultrasound-induced changes in blood-brain barrier permeability, but the comparison in this study examined distribution after direct injection. Reading it as a demonstration of opening the barrier to let a drug through would misconstrue what the device improved.
Combining a thin injection tube with an ultrasound stimulator
LUCID joins local-delivery and ultrasound technologies that Dagdeviren and colleagues had developed separately. Comparing the device roles and the presence of a fluid channel across MIT's official announcements at each point shows how the work led to this integration.
| Official announcement | Device role | Relation to this work |
|---|---|---|
| January 24, 2018 | Local delivery to the rat brain using multiple thin tubes and a pump | A route for getting substances into the brain |
| June 4, 2024 | Ultrasound neural stimulation of the mouse brain with ImPULS | Expressed interest in adding microfluidic channels |
| October 26, 2026 | LUCID integrates an injection tube with 24 ultrasound transducers | A design to aid distribution in tissue after injection |
The 2018 work was local delivery into the brain, and the 2024 work was neural stimulation with local ultrasound. LUCID integrates the injection tube and ultrasound transducers and moves toward a design that widens distribution within tissue after injection.
The 2024 ImPULS announcement had expressed interest in eventually adding microfluidic channels. In 2026, the team used channels and ultrasound in a single device and examined how far the injected substance spread. However, the animal experiments with the earlier devices had different purposes and are not follow-up replications of LUCID. What this timeline lets us compare is device roles and configuration. It does not show which performs better therapeutically, or that clinical demonstration has advanced step by step.
How far substances spread in gel and in mice
The first comparison used a tissue-mimicking gel made to resemble human tissue. According to MIT, when ultrasound was used, drug molecules distributed more widely than without it. This was an experiment in which substances were introduced into a real material, not a computer simulation.
In the subsequent mouse experiment, the substance injected was an MRI contrast agent. MIT explains that when the ultrasound array was activated, the contrast agent distributed more widely in the brain than when it was not. The target moved from gel testing to an animal brain, and the difference in distribution was examined. However, this is neither a result of treating a tumor with an anticancer drug nor a trial confirming efficacy in humans.
The team varied the presence of ultrasound for comparison, which distinguishes it from a study that merely observes naturally occurring differences in distribution to look for associations. Even so, the paper's main text and supplementary information could not be checked, so the statistical design cannot be directly evaluated. From the university's description, the size of ultrasound's causal effect, and how much each mechanism contributed to distribution, cannot be determined.
In the published materials, the number and strain of mice, as well as the allocation method, are not stated. The injection volume, and the ultrasound intensity and duration, are not given either, so there is not enough information to replicate under the same conditions. Absolute values of distribution volume and variability are also missing from the materials obtained, and for the gel tests the sample size and quantitative values are unknown.
What can be reported at this stage, therefore, is the qualitative comparison the university reported: distribution was wider with ultrasound. No figures are available for judging how many times wider it became or how consistently this occurred. The 24 transducers and the roughly 100-microliter reservoir are device specifications, not numbers indicating the scale or effect size of the tests.
Reducing recurrence will require validating drug efficacy and long-term safety
The motivation behind the research is the recurrence of glioblastoma, a malignant brain tumor. Even when a tumor is surgically removed, it is difficult to eliminate every cancer cell. MIT raises the possibility of using this device as a way to deliver drugs to cells remaining after resection.
Implantable materials that place drugs in the brain are already being studied, and in MIT's explanation such materials rely on passive diffusion of the drug. LUCID is designed to add ultrasound to help distribution into the surrounding tissue. But a wider reach for the substance and sufficient drug effect on tumor cells are separate matters that need to be verified independently.
The results confirmed so far do not include a reduction in recurrence rate or an extension of survival. Killing of cancer cells and reduced systemic toxicity have also not been confirmed as measurement results in this work. Ornellas explains that enhancing distribution locally could reduce toxicity to off-target organs, but this is an expectation for future treatment.
For a human version, the team described a concept in which the injection tube and ultrasound transducers sit inside the skull, with the battery and electronics placed outside it. Immunotherapy and delivery of nucleic acids are named as candidate uses, but they are not demonstrated indications.
Within the scope of the public materials searched, no independent replication of LUCID by another team could be found. BrightSurf's article is also a reprint of MIT's announcement; it supplements the publication date and DOI as bibliographic information, but does not count as an independent reproduction.
The next question is whether the device can deliver enough drug to the necessary area using actual drugs and disease models. Dosing conditions must be set, and the effects on tissue and the drug's efficacy must be confirmed with long-term use. Reaching that point would move the research from a device that distributes substances widely to one whose therapeutic benefit can be evaluated.