She is putting an entire imaging modality onto a chip - and reading the body's faint magnetic fields at room temperature.
To measure the brain today, you usually need a superconductor chilled to roughly minus 270 degrees Celsius, a shielded room, and a machine the size of a small car. Nishita Deka thinks all of that belongs in a chip you could wear.
Every time a muscle twitches or a cluster of neurons fires, the body throws off a magnetic field. It is real, it is information-dense, and for most of the last century it has been almost impossible to read without exotic hardware. Nishita Deka runs a company built on the wager that this is about to change.
Sonera, the Berkeley startup she co-founded and leads as CEO, makes chip-scale biomagnetic sensors. The first product is called the S1, and its job is to detect the magnetic signature of muscle activity - a technique the company helped name: magnetomyography. The same physics points toward something bigger: portable, high-fidelity sensing of brain activity that does not require a cryogenic plant in the basement.
Her framing is disarmingly plain. "We are trying to detect brain activity using cheaper, faster methods that are still high-performance," she has said. The conventional tools, in her words, "are fairly rudimentary and limited in either signal quality or usability." Sonera's sensor is meant to be, as she puts it, cheap, small, low-power, and light enough to be mass-produced.
The near-term uses read like a list of things that have been waiting on better hardware: gesture control, augmented and mixed reality, prosthetics, neurorehabilitation, and silent-speech interfaces where the system reads the muscles of speech before a sound is ever made. The far-term use is the one that makes neuroscientists lean in - genuinely portable brain imaging.
Detecting brain activity should be as easy as it is to detect heart rate today.
Say that five times fast. Underneath the mouthful is an elegant trick: Sonera's sensors exploit the interaction between thin magnetic films and high-frequency sound waves to pick up magnetic fields far too faint for ordinary electronics. No supercooling. No shielded room.
Firing neurons and contracting muscles produce tiny magnetic fields.
A magnetic thin film, driven by sound waves, shifts its resonance in response to those fields.
That shift becomes a clean electrical signal - on solid-state silicon, at room temperature.
Traditional magnetoencephalography (MEG) buys its sensitivity with cryogenics and shielding. Sonera is trying to keep the resolution and throw out the cold.
It could change how MEG is used entirely and make it much more accessible.
Before high school, Nishita Deka wanted to be a pediatrician. Then a physics teacher got hold of her imagination, and the plan quietly fell apart in the best way. At USC she did applied-physics research in an optics group under advisor Andrea Armani. At UC Berkeley she went deep into semiconductor devices, working on nanoscale structures built from two-dimensional materials and high-throughput fabrication.
Sonera did not begin as a pitch deck. It began as a friendship. She met co-founder Dominic Labanowski while both were grinding through PhDs in electrical engineering and computer sciences at Berkeley. He was working on acoustically-driven ferromagnetic resonance and multiferroic materials; her background in devices and fabrication slotted in alongside it. They built the partnership on shared values first - and a conviction that healthcare could be improved through better instruments. With the support of Labanowski's PhD advisor, Sayeef Salahuddin, an IEEE Fellow, they launched the company in 2018.
There was a personal current running underneath the engineering, too. Deka has spoken about her own years-long struggle to get a PCOS diagnosis, and how that experience convinced her that "easy access to really high-resolution imaging could improve people's understanding of their own health, a lot." The mission was not abstract. She had lived the gap she wanted to close.
The same year they founded the company, Deka became an Activate Fellow at Cyclotron Road, the deep-tech program housed at Lawrence Berkeley National Laboratory - two years of runway to turn lab physics into something the world could actually buy.
"Developing new hardware takes a lot of time, even just to demonstrate basic capabilities." No demo-day shortcut bends the physics of fabrication.
Biomagnetic fields are vanishingly small. The whole company is, in a sense, an argument about how to hear a whisper without a soundproof room.
Magnetomyography barely existed as a named modality. Sonera co-authored the preprint that helped put it on the map in 2024.
Sonera's 2023 seed round was led by Amplify Partners, with co-investors including Abstract Ventures, Spark Capital, Material Impact, and Boom Capital. Add government awards - including a National Science Foundation grant - and a partnership with the U.S. Air Force Research Laboratory, and the running total reaches roughly $20 million.
Seeing our team's ideas go from hypotheses to reality is incredibly satisfying.
"We are trying to detect brain activity using cheaper, faster methods that are still high-performance."
"Current methods for measuring brain activity are fairly rudimentary and limited in either signal quality or usability."
"By making MEG technology portable, more people will be able to access better diagnostics."
"One big challenge is that developing new hardware takes a lot of time, even just to demonstrate basic capabilities."
Make reading the nervous system as ordinary as taking a pulse. Put high-fidelity neural and muscular data into something portable, wearable, and affordable enough that it stops being a hospital privilege and starts being a fact of everyday life.