Researchers at Caltech have developed a minimally invasive brain-machine interface that uses functional ultrasound to decode movement intentions with unprecedented precision. Unlike current BMI systems requiring electrodes implanted directly into brain tissue, this approach captures neural activity through a small skull window, dramatically reducing surgical risk while maintaining high-resolution capabilities.

The team achieved 100 micrometer resolution by measuring blood flow dynamics in the posterior parietal cortex, a region critical for movement planning. Their method produces detail approaching traditional electrophysiology techniques while avoiding the need to penetrate the dura membrane. This addresses a major barrier limiting BMI adoption: the extensive brain surgery currently required for implantation.

Testing with non-human primates demonstrated real-time decoding accuracy, predicting eye and arm movements within seconds of intention formation. Machine learning algorithms successfully correlated ultrasound imaging patterns with specific behavioral tasks, validating the approach against established electrophysiology benchmarks.

Published in Neuron on March 22, the research opens pathways for human trials with patients who have existing skull openings from traumatic injury. This advancement could expand BMI accessibility beyond the small population currently eligible for invasive procedures.