A headband that aims sound at the deep brain
Most founders chase the surface. Rajiv Mahadevan went for the middle of the skull. Attune Neurosciences, the company he co-founded in 2019 and has run as CEO since 2020, is building what it bluntly calls the world's first wearable, non-invasive ultrasound system for deep brain therapy. No surgery. No implant. No drug. Just low-intensity focused ultrasound, steered through bone to a target a few millimeters wide, with the path planned from a person's own brain scan.
The idea is almost rude in its simplicity. Sound waves pass through the head every time a doctor runs an ultrasound. Attune's wager is that those same waves, focused tightly enough and aimed precisely enough, can nudge specific deep brain structures - and that the whole apparatus can shrink down to something you wear like a hat instead of something a neurosurgeon installs. The early device, a headband-shaped investigational system, also listens while it works: on-board electrodes read brain rhythms and an accelerometer tracks head movement, so the stimulation can adapt in real time rather than fire blind.
It is a strange and specific bet, and it is exactly the kind Mahadevan keeps making. His resume is not a straight line up a ladder. It is a series of jumps across the frontier where new biology becomes something you can actually ship - and each jump lands a little closer to the brain.
DNA, then drugs, then the brain
Start at Berkeley. Mahadevan studied Molecular & Cell Biology at UC Berkeley in the mid-1990s, the unglamorous bench-science apprenticeship that teaches you how living systems actually behave. Then Stanford's Graduate School of Business, where he picked up the second language he would need: how a discovery turns into a company, a balance sheet, an exit.
With both halves in hand, he went to work at the seam. There were business development roles at Novo Nordisk and at 23andMe - the consumer-genetics company that, for a moment, made it normal to spit in a tube and read your own code. He co-founded and ran Applied Immunology as CEO, then watched it get acquired by Precision Medicine Group, where he stayed on as a Managing Director from 2016 to 2020. He sat on the board of Bionure. He became an investment partner at Aliso Ventures, learning the table from the other side. In 2021 he co-founded Tactile Therapeutics. The pattern is hard to miss: he keeps showing up wherever a piece of frontier biology is one good operator away from becoming a real product.
Attune is where the instinct points its sharpest. Immunology and genomics taught him to read the body. Neuromodulation is the harder, stranger move - not just reading the system, but reaching in to change it, and doing it gently enough that you can wear the device to bed.
The skull is the problem - and the point
Aiming ultrasound at the deep brain sounds easy until you remember what's in the way: a curved, uneven slab of bone that bends and scatters sound differently for every single person. Point the same beam at two different heads and it lands in two different places. That is why Attune leans on imaging - the company describes MRI-guided precision and adaptive positioning, using a person's own scan to compute how the waves will travel and where they will converge.
The company's patent work centers on head-wearable devices for positioning ultrasound transducers, and on transducers that can both stimulate and sense. The first targets are about sleep and wakefulness and human performance - the foundational rhythms that govern how a brain rests and fires. Get the wearable form factor and the adaptive aim right there, and the same platform can travel deeper into harder problems later. That is the long game: build the hardest version of an easy-sounding promise, then let it generalize.
Wear it, don't implant it
A headband-shaped system instead of a surgical procedure. The whole bet is that deep-brain neuromodulation can leave the operating room and join the things you simply put on.
Aimed by your own scan
MRI-guided targeting and adaptive positioning compute how sound bends through one specific skull, so the focal point lands where it's supposed to - not where an average skull would send it.
It listens while it works
On-device electrodes and motion sensing let the system read brain rhythms and adapt the stimulation in real time, instead of broadcasting a fixed signal and hoping.
Sand Hill meets SOCOM
The funding roster mixes venture investors and the NIH with the U.S. Special Operations Command - an unusual coalition that says a lot about who's interested in non-invasive performance tools.