Researchers at Cornell University have developed a neural implant smaller than a grain of salt — one that can, for the first time, record electrical activity from the brain during magnetic resonance imaging (MRI) scans. The achievement represents a landmark in neuroscience research and brain-computer interface (BCI) technology, opening possibilities that were considered technically impossible until recently.
The core problem that motivated this research was straightforward: traditional neural implants are built with metallic components — tungsten wires, titanium housings, and conductive alloys — that make them fundamentally incompatible with the MRI environment. The intense magnetic field causes heating in these implants, while conductive materials introduce artifacts that degrade image quality. In clinical terms, no patient with a conventional neural implant can safely undergo an MRI scan — a significant limitation that affects both diagnosis and longitudinal disease monitoring.
The Solution: Radical Miniaturization and MRI Compatibility
The Cornell team’s approach circumvented this problem elegantly: by eliminating metallic components from the implant entirely. Rather than using conventional wires and circuits, the device operates using piezoelectric materials — substances that generate electrical current when subjected to mechanical pressure, such as that exerted by ultrasound waves. This means the implant can be wirelessly powered and can transmit data via ultrasonic waves, without any metal antenna that would interfere with MRI’s magnetic field.
The result is a device with a volume smaller than one cubic millimeter — tinier than a grain of rice and comparable to a grain of table salt — capable of recording electrical signals from individual neurons or small neuronal clusters. Captured data is transmitted outside the skull through ultrasonic transducers positioned on the scalp, eliminating the need for transcutaneous cables that increase infection risk in chronic implants.
Why MRI Compatibility Is a Game-Changer
The ability to record brain activity simultaneously with an MRI scan is not a minor technical detail. Functional MRI (fMRI) measures brain activity indirectly through the BOLD signal (Blood Oxygen Level-Dependent) — a hemodynamic measure with temporal resolution limited to seconds. Direct electrophysiological recordings from neural implants, by contrast, capture electrical variations in milliseconds — far closer to the true speed of neuronal processing.
Combining the two modalities — the spatial resolution of MRI with the temporal resolution of electrophysiology — has long been a theoretical objective in neuroscience. A MRI-compatible implant now makes it possible to study, for example, how neural networks in epilepsy patients behave during a seizure, or how specific regions of the cortex respond to sensory stimuli with a precision that fMRI alone could never deliver.
For radiology, the immediate impact is indirect but relevant. Patients with neural implants were previously automatically excluded from MRI protocols — a limitation affecting both diagnosis and follow-up of neurological diseases. As MRI-compatible devices become available, radiologists will increasingly encounter patients with next-generation implants whose scanning is no longer contraindicated.
Brain-Computer Interfaces: The Broader Horizon
Cornell’s implant development fits within a broader context of growing commercial and scientific interest in brain-computer interfaces. Companies like Neuralink and academic groups worldwide are developing devices enabling people with paralysis to control computers, robotic limbs, or communication systems directly through neural activity.
Radical miniaturization — taken to a new extreme by the Cornell team — is critical for clinical application of BCI technology. Smaller implants cause less tissue damage during insertion, generate less chronic inflammatory response, and can be positioned with greater precision in specific cortical regions. The elimination of transcutaneous cables, in particular, is a fundamental step toward long-term implants in real clinical use.
The field still faces significant stability challenges — the quality of electrical recordings tends to degrade over time as scar tissue encapsulates the implant. But MRI compatibility opens a new scientific avenue: the ability to monitor chronically implanted subjects during imaging scans, and to understand how the brain adapts to the presence of artificial devices over time.
Implications for Radiology and Clinical Practice
In the short term, the most immediate application for this type of implant is scientific: recording brain activity in animal models or, eventually, in human volunteers during MRI sessions. In the medium term, the technology could reshape neurological imaging protocols — generating new demand for MRI in patients carrying next-generation implants.
For radiology departments already preparing for the expansion of BCI technologies, MRI compatibility in new devices simplifies pre-scan screening and eliminates the need for alternative imaging modalities (such as CT) in implanted patients. As MRI remains the dominant modality in neuroimaging — and with ongoing challenges around scanner availability and operational costs — the arrival of MRI-safe implants represents a genuine expansion of diagnostic coverage for neurological patients.
The Cornell research is still in pre-clinical stages, but the proof of concept — a sub-millimeter, metal-free, MRI-compatible neural implant — opens a new chapter in translational neuroscience and sets a direction for the next generation of BCI technology.
Source: Inovação Tecnológica
