University of Michigan achieves first human brain recording with wireless implant

Highly anticipated: As the brain-computer interface technology field transitions from experimental demonstrations to practical clinical applications, the University of Michigan has just achieved a milestone that could make high-performance, wireless BCIs a reality for people living with paralysis, speech loss, and other challenging conditions. The successful test of a fully implantable device in a human patient signals a new era in neurotechnology, where restoring lost functions through brain-to-computer communication is moving closer to everyday clinical use.

In a significant advance for brain-computer interface (BCI) technology, a University of Michigan research team has achieved the first in-human recording using Paradromics’ Connexus device – a wireless, fully implantable BCI designed to restore communication and movement for people living with severe neurological conditions. The procedure took place on May 14, 2025, during epilepsy surgery, where the device was temporarily placed on the patient’s temporal lobe, an area essential for processing sound and memory. This opportunity allowed the team to safely test the device’s ability to capture neural signals without adding risk to the patient, as the surgery already required access to the brain.

The Connexus stands out for its compact size – smaller than a dime – and its high-density array of 421 microelectrodes, each thinner than a human hair. Unlike many earlier BCIs, which often relied on fewer electrodes and required external wires, Connexus is engineered to be fully implantable. The device collects electrical signals from individual neurons, transmitting them via a thin lead to a transceiver implanted in the chest. From there, the data is sent wirelessly to an external computer, where artificial intelligence algorithms interpret the patterns and translate them into actions, such as moving a cursor or generating synthesized speech.

This high-resolution approach is a notable departure from other BCI systems that monitor groups of neurons from the brain’s surface or even from within blood vessels. By targeting individual neurons, Connexus aims to deliver more precise and nuanced decoding of brain signals, which is crucial for restoring natural communication speeds and improving device performance. For example, recent advances in BCI technology have enabled the decoding of intended speech at rates approaching 78 words per minute; however, the goal is to reach the pace of natural conversation, which is approximately 130 words per minute.

The surgical team used a specialized, EpiPen-like instrument to implant the device, demonstrating that the procedure can be performed with tools familiar to neurosurgeons worldwide. This is expected to help facilitate broader adoption and safer clinical practices as the technology matures.

Paradromics’ focus is not only on technical innovation but also on durability and long-term safety. Previous animal studies have shown that the device can maintain stable signal quality for over two and a half years without degradation. The company is now preparing for a clinical trial, pending regulatory approval, that will enroll ten participants and monitor the device’s safety, performance, and impact on patients’ lives for a year.

The potential applications of Connexus extend beyond restoring speech and movement. By decoding neural signals at the level of individual neurons, the technology could one day help address mental health conditions or chronic pain by interpreting mood or discomfort directly from brain activity.

This broader vision reflects a growing momentum in the BCI field, which has attracted more than $2 billion in investment and is being challenged by companies like Neuralink and Synchron, each pursuing distinct technological strategies.

At the University of Michigan, this breakthrough is part of a larger effort to develop next-generation BCIs that are more reliable, less invasive, and capable of delivering long-term benefits to patients. The team’s work builds on decades of BCI research, moving beyond the limitations of earlier devices, such as the Utah array, which required external connectors and could cause tissue damage over time.