Paradromics has secured US Food and Drug Administration (FDA) approval to begin a human trial for its Connexus brain implant system. The device is intended to restore speech for individuals with severe motor impairments.
Speech Restoration Trial Approved
The trial is set to start in early 2025 with two participants. Its primary goals are to assess the Connexus implant’s long-term safety and its effectiveness in facilitating synthesised speech and text communication. This pivotal trial represents a significant step forward for Paradromics and the broader field of assistive technology. The FDA’s green light signifies a rigorous review process, indicating confidence in the preclinical data and the safety protocols established by the company. For individuals living with conditions that rob them of their ability to communicate verbally, such as advanced amyotrophic lateral sclerosis (ALS), locked-in syndrome, or severe stroke-related paralysis, this trial offers a tangible beacon of hope. The focus on both safety and efficacy underscores the dual challenges in developing such advanced neurotechnology: ensuring patient well-being while delivering a functional and meaningful improvement in quality of life.
Decoding Brain Signals
Paradromics, headquartered in Austin, Texas, aims to provide a digital voice for individuals unable to speak due to conditions like amyotrophic lateral sclerosis (ALS) or stroke. The system decodes brain signals associated with the intent to speak, translating them into commands for communication devices. This sophisticated process relies on advanced machine learning algorithms and neural signal processing. The core innovation lies in the ability of the Connexus system to interpret the subtle electrical activity within the brain that corresponds to the desire to form words and sentences. Unlike simpler brain-computer interfaces (BCIs) that might detect broad motor commands, Connexus is designed to capture the nuanced neural patterns indicative of speech intention. This requires highly sensitive electrodes capable of recording from individual neurons, offering a granular view of brain activity.
Trial participants will attempt to vocalise words and sentences. The Connexus system will learn to recognise the specific brain patterns generated during these attempts. These patterns are then translated into text displayed on a screen. The company plans to use AI to generate a voice clone from existing recordings of the participant’s voice to vocalise the synthesised speech. This AI-driven voice cloning is a crucial element, aiming to restore not just the ability to communicate, but to do so with a familiar and personal voice. The system’s ability to adapt and learn from each participant’s unique neural signatures is key to its potential success. The iterative process of decoding, translating, and vocalising will be refined over the course of the trial, with the ultimate goal of achieving naturalistic communication.
AI Voice Cloning
Matt Angle, CEO and founder of Paradromics, stated the system is projected to enable communication speeds of up to 60 words per minute, supporting sustained conversation. This rate is comparable to speeds reported in prior brain-computer interface (BCI) trials by academic research groups. Typical human speech ranges from 120 to 150 words per minute. While 60 words per minute is a significant improvement over current assistive communication methods for many, it still represents a substantial gap from natural speech. However, the ability to engage in sustained conversation, even at this reduced rate, could profoundly impact social interaction, professional life, and personal relationships for those affected by severe speech impairments. The company’s focus on achieving this speed suggests a pragmatic approach, balancing technological feasibility with user needs. The comparison to academic research highlights the competitive landscape and the ongoing drive for innovation in the BCI space.
The Connexus implant is a small metal disk containing 421 microwire electrodes. These electrodes are designed to rest within brain tissue and record signals from individual neurons. The device requires surgical implantation. The technical specifications of the implant are critical for its performance. The high density of electrodes allows for a rich dataset of neural activity to be captured, which is essential for training the AI algorithms. The choice of microwire electrodes suggests a focus on precise neural recording. The requirement for surgical implantation means that the benefits of the Connexus system must demonstrably outweigh the risks and invasiveness associated with neurosurgery. This will be a key consideration for potential recipients and their medical teams.
Earlier in 2024, Paradromics performed a brief, 10-minute implantation of its device in a patient already undergoing brain surgery. This procedure tested the surgical insertion and removal process; the device was not used for speech restoration at that time. The upcoming trial will involve long-term implantation. This prior, short-term procedure served as a vital proof-of-concept for the surgical methodology, allowing the team to refine techniques and gather preliminary data on device integration with brain tissue without the pressure of immediate functional assessment. The transition to long-term implantation in the upcoming trial is a critical step, enabling the system to be used continuously and for the AI to achieve optimal learning and performance over an extended period. The success of this long-term implantation will be a key indicator of the system’s viability for widespread clinical use.
