As previously mentioned, the sort of hardware we’re discussing can be manufactured relatively inexpensively and operates quickly and at very low power. Add it all up, and we foresaw a potential imaging solution that addresses many of the most deeply entrenched problems in medical imaging. The icing on the cake was the possibility that appropriately crafted neural networks could "learn" the dielectric signature of normal and pathologic tissue states, and here the motivation for tuning such a system for detection of strokes became obvious.

HCB News: The artificial neural networks are doing the heavy lifting, but none of this would be possible without the hardware you developed in partnership with NYU’s LaGuardia Studio, right?
SD and LA: That’s correct; With the help of LaGuardia Studio we’ve designed and fabricated a fully-fledged prototype of the helmet. Specifically, the mechanical structure was built to be lightweight, comfortable for patient scanning, and seamlessly interface with all the lab-fabricated electronics necessary for scanning. To date, we’ve scanned a few dozen healthy human subjects, having now documented the potential for anatomically realistic brain imaging in 3D from the system, with true MR-like contrast.
We’ve also developed a dedicated experimental benchtop framework for further advancing the core technology and more intensively studying the properties of the technology in realistic human brain phantoms, which we recently reported in the journal Nature: Communications Engineering.

By 3D printing geometrically varied hemorrhage phantoms with the electrical properties of blood, we can navigate the hemorrhage throughout the head model and rigorously test the performance for detection of hemorrhage, for classification of hemorrhage subtypes, sizes, and shapes, and for precise localization in 3D space. This allows us to optimize the acquisition protocol and neural network in order to streamline it, making it as fast, lean, and as accurate as possible.

HCB News: Where do you see technology like this making the biggest impact?
SD and LA: Any of the challenges described thus far can be thought of as being amplified outside the wealthiest urban areas and are overwhelmingly the burden of the most economically disadvantaged. In less affluent and less industrialized parts of the world there may be literally no access whatsoever for vast stretches across an entire country.

Imagine a system deployable with the ease of something like a defibrillator (forgive the very loose analogy) — a true grab-and-go piece of technology with a footprint small enough and light enough for ubiquitous availability, and at a cost that doesn’t further perpetuate massive imbalances in global health.

More Magazine Features

Stroke Homepage