Paradromics

A little birdie told her. What could neural data tell us? Dr. Julie Elie recently decoded birdsong – specifically the language of zebra finches – by identifying the core calls that make up their vocabulary and what each one means. For more than a decade, she recorded the sounds the birds made and classified them by context and the bird that made them. She then used machine learning to analyze how information was encoded in each call. Finally, she confirmed her work by testing whether the finches themselves agreed with her classifications. For example, in one test, birds tapped a button to hear calls from their repertoire, with some calls followed by a seed reward. When they made mistakes, they tended to confuse calls that shared the same meaning rather than the same sound, suggesting they grasped what the calls actually meant. Her method is strikingly close to how neural decoding with an intracortical brain-computer interface (BCI) works. When someone attempts to speak, neurons in the motor cortex produce electrical signals associated with the intended word. Machine learning models then analyze the neural patterns recorded by the BCI to identify what is being expressed. And the person confirms, in real time, whether or not the intended meaning came through (no bird seed required). Both ask the same fundamental question: What is being communicated, and how do we learn to understand it? Decoding the brain is a vital part of the engineering behind restoring lost functionality. However, it may also teach us something larger: how thought itself is organized, providing us with a window into the brain unlike ever before. Read more about Dr. Elie’s decoding process at the link in the caption.

What is needed to shift next-generation robotics from a promising prototype to something people with limb loss can access, afford, and rely on? This #SXSW2027, Phantom Neuro is proposing a panel, "Concept to Consumer: Scaling Access to Next-Gen Robotics," that seeks to shed light on the answer. The event – now open for community voting – will bring together Phantom Neuro CEO Connor Glass, Paradromics CEO Matt Angle, Aaron Holm of Wiggle Your Toes, patient and technology advocate Katie Meyer, and Ottobock. If this is a session you'd want to see on the SXSW stage this coming spring, we'd be grateful for your vote. Cast yours at the link in the caption by searching the panel title.

Long before we understand the rules of language, the brain learns how sounds fit together. Follow a single word, “cat,” as an intracortical brain-computer interface, or BCI, creates a new path from intention to expression. When you attempt to speak, neurons in the motor cortex produce electrical signals associated with the intended word. Machine learning models analyze the neural patterns and identify the likely building-block sounds: /k/, /a/, /t/. These sounds are called phonemes, and the English language contains about 40 of them. Language models then use those sounds and their context to predict the intended word: cat. All of this is engineered to happen in the space of a breath. Natural conversation moves at approximately 160 words per minute, and a speech BCI must keep pace with the rhythm and flow of real communication. Our latest blog explores how BCIs may use the natural building blocks of speech to turn attempted speech into expression. Read more at the link in the comments.

A brain-computer interface (BCI) may enable many different capabilities, but they all depend on one essential standard: it must work reliably for the people who depend on it, for years to come. This comes down to three factors, working together in harmony. Safety comes first, always. Performance is tied to capturing neural signals at the highest possible resolution and turning them into communication that conveys the user’s intent at a natural speed as opposed to feeling stilted or robotic. And durability means a device is engineered to last for years without a replacement. Without all three, BCIs will remain experimental. With them, they can become technology people trust and rely on for years. In our blog, Chief Scientific Officer Vikash Gilja, Ph.D, and Chief Medical Officer Stephen Ryu, M.D., break down the design choices behind the Connexus® BCI and highlight why these three pillars should set the standard for the field. Read on at the link in the comments.

Ten years ago, tech writer Camille Tuutti decided to get an RFID chip implanted in her hand at a party in Stockholm, just because she thought it was cool. It’s roughly the size of a grain of rice, and if you scan her hand with a phone, her contact information will pop up. Her conversation with Paradromics Founder and CEO Matt Angle covered how the field of implantable tech has come from a novelty device you could get at a party to a brain-computer interface (BCI) designed to restore lost functionality. They also confronted a challenge that’s yet to be solved: reimbursement. Without established billing codes, first-in-class BCIs can wait a year or more for Medicare coverage, and this gap is what stands between the technology and the people it’s built to serve. While new efforts like the FDA-CMS RAPID program are currently underway to help streamline the process, there’s still much work – and collaboration – to be done. Head to the link in the comments to read the full article.

When someone uses a brain-computer interface (BCI) to speak and thoughts transform seamlessly into words, it can almost seem like magic. But in reality, a speech BCI is designed to work by reading intended actions directly from the brain.    This occurs in three stages: recording the neural signals produced when someone attempts to speak, decoding those signals into intended words (often by first identifying phonemes, the building-block sounds of language), and outputting a desired action like text or synthesized speech, the latter of which can even be in the user’s own voice if they conducted voice banking.   For the hundreds of thousands of people living with ALS and other neurodegenerative conditions, the loss of speech also means the loss of independence, interaction, and the individual voice needed to advocate for oneself. Restoring even part of it – especially at the speed and nuance of real conversation – would play a central part in giving that back.   Our latest blog explores how speech decoding actually works, stage by stage. Read more at the link in the comments.

On World Brain Day, we want to take a moment to recognize the people at the center of our work: neurosurgeons and neurologists. Restoring communication for people who have lost the ability to speak is an innately human endeavor. Building technology that puts clinicians at the heart of it is one of our proudest achievements. To David Brandman, MD, PhD, Dr. Matthew Willsey, and the neurological and clinical teams behind the Connect-One Clinical Study: thank you for your commitment, your care, and your continued partnership. The people this technology is designed to serve have a brighter future because of you.

Longevity and safety are two necessary pillars of a successful brain-computer interface (BCI). A device designed to last a lifetime must prove it remains both safe and functional for years, and a recently published study from our APEX partners Jennifer Collinger and Robert Gaunt demonstrates just that. Their work represents the longest-running safety study of an intracortical micro-stimulation (ICMS) system, a technology that aims to restore a sense of touch by delivering small electrical pulses to the brain’s sensory cortex. Over a combined implant duration of 27 years across five participants with spinal cord injury, the system delivered more than 168 million pulses with no serious adverse events. The majority of electrodes continued to reliably evoke sensations of touch localized to the hand, some for as long as 10 years. Feedback from the hand is essential for coordinated movement and control, and evidence that these systems can safely and durably deliver that feedback over a decade marks a significant step toward building BCIs that patients can depend on for the long haul. Studies like this build the collective foundation supporting the safety and longevity of intracortical BCI systems. Read the study in full in the comments below.

Even when someone can no longer move the muscles required for speech, the part of the brain that plans those movements can remain active. It may still be sending the signals for ideas the body can no longer express. That is why Paradromics developed the Connexus® BCI, the highest-data-rate brain-computer interface in the world. By recording and decoding neural activity from individual neurons and processing that data in real time, the device is designed to restore natural human speech for those who have lost it. We cannot yet cure every disease. But we can help restore one of the most human capabilities of all: not just communication, but the ability to express who we are.

A brain-computer interface (BCI) succeeds only if its design can deliver purposeful, reliable functionality for the people it is intended to serve. Achieving that requires deliberate engineering choices built around three foundational pillars: safety, performance, and durability. Decades of experience with implantable medical devices and neurosurgery have demonstrated that long-term implantation can be performed safely. For any BCI designed to remain in the body for years, however, safety must always be the first non-negotiable pillar. Performance determines whether a BCI can capture neural intent at the highest possible resolution and translate it into meaningful improvements in communication and mobility. Durability ensures that the device can continue delivering reliable function for a decade or longer without requiring replacement. Only by delivering all three can BCIs become trusted, clinically viable technologies capable of reaching the people who depend on them. In our latest blog, Chief Scientific Officer Vikash Gilja, PhD, and Chief Medical Officer Stephen Ryu, MD, explain the engineering design choices behind the Connexus® BCI, why safety, performance, and durability guide our work, and why they should set the standard for BCI development. Head to the link in the comments to learn more.