Unveiling the Future of Brain Imaging: Ultrasound Technology in Focus
A few years ago, an article was published that blew us away. The idea was that we could decode what someone was looking at simply from their brain activity. This remarkable concept opened up a new frontier in neuroscience, hinting at a future where telepathic communication could become a reality. However, the technology of the time required an MRI machine, which is not feasible for everyday use as it cannot be worn on the head.
The Hardware Challenge in Mental Interfaces
The field of mental interfaces has long faced a significant bottleneck: the hardware. On one end of the spectrum, there’s the invasive approach of drilling a hole in your skull to insert electrodes. On the other, there’s the non-invasive but limited method of using EEG to record brain activity from outside the head, resulting in blurry images. The quest for a non-invasive yet detailed solution has been ongoing, and now, a promising new technology has emerged.
Introducing Ultrasound-Based Brain Imaging
We have developed a new type of hardware that offers MRI-level detail without the need for invasive procedures. This groundbreaking technology is based on ultrasound, leveraging the connection between the vascular system and neurons. When neurons fire, they require more blood, and we use ultrasound waves to map this blood flow and volume throughout the brain, all through the intact skull.
Achieving Detailed and Comprehensive Brain Imaging
For a general-purpose mental interface, two key requirements must be met: a wide field of view and high resolution. Traditional modalities like EEG and MEG provide a large field of view but lack detail. Neurovascular ultrasound, akin to MRI, meets both criteria. It enables the recording of a million independent pixels across the brain, with each pixel measuring less than a millimeter. Until now, achieving such detail required removing the skull, but the challenge lies in maintaining the skull intact while capturing such precise images.
First Light: A Milestone in Brain Imaging
Today, we are thrilled to announce a milestone: the most detailed vascular image of a living human brain ever captured through the skull. This achievement marks the world’s first 3D image of ultrasound localization microscopy in a human brain through an intact skull, offering 100 times higher volumetric resolution than comparable CT scanning.
We anticipate numerous applications of transcranial microbubble imaging beyond our current focus. Diseases like stroke, Alzheimer’s, and head trauma each leave unique vascular signatures at scales that CT and MRI cannot resolve. We believe imaging at this resolution will greatly enhance diagnostic capabilities, which is why we are open-sourcing our entire pipeline and dataset.
Microbubble Processing Pipeline
Microbubbles allow us to surpass the diffraction limit. Normally, ultrasound cannot separate two objects closer than one wavelength, but microbubbles enable us to pinpoint their centers with subpixel accuracy. By injecting small bubbles that don’t overlap, we can map their positions with precision far beyond traditional limits. As these microbubbles flow through the vasculature, we amass millions of positions to create an image with detail finer than the wavelength itself.
The bubbles, composed of sulfur hexafluoride encapsulated in lipid shells, are an FDA-approved contrast agent. During a four-minute acquisition, we continuously infuse these bubbles. Their distinct acoustic impedance enhances the signal, allowing for super-resolution imaging. By tracking the bubbles in 3D, we can trace blood flow through the living microvasculature.
Towards Contrast-Free Neurovascular Imaging
Our contrast-enhanced results are a significant step forward, providing a reliable image of vascular details through an intact skull. Our ultimate goal, however, is to achieve non-contrast neurovascular imaging of the brain. Two trends fuel our confidence in reaching this objective. First, the hardware: ultrasound machines, once costly and cumbersome, are now the size and price of a smartphone thanks to innovations by companies like Butterfly. Second, data: while imaging without contrast presents challenges due to weaker signals from red blood cells, advances in machine learning offer hope. By training on extensive datasets, we can potentially recover signals that current methods miss.
We are actively building what we believe to be the world’s largest neurovascular disease dataset using ultrasound technology. We are excited to see what the future holds and are eager to share our progress. For more information, visit our source here.
“`
