This man with ALS is “the first power user” of a brain implant that lets him speak

2026-06-15 · Source: MIT Technology Review · Field: Health & Wellbeing — Medical Devices & Health Technology, Artificial Intelligence & Machine Learning, Clinical Care & Medical Practice · Depth: Novice, medium

Summary

Casey Harrell, a 45-year-old man with amyotrophic lateral sclerosis (ALS) and paralysis, has been using a brain-computer interface (BCI) with embedded electrodes for almost three years. After a five-hour operation in July 2023 to implant four arrays of 64 electrodes each, Harrell began using the device to "speak" sentences. Within 22.6 months, he logged over 3,800 hours of independent use, becoming the "first power user" of a speech BCI. The system decodes brain activity from the speech motor cortex into 39 American English phonemes, then into words. Initially, it achieved 99.6% accuracy with a 50-word vocabulary, later expanding to 125,000 words with 97.5% accuracy, and is now 99% accurate. Harrell's care partner can now connect him, enabling independent use for communication, web surfing, and his job. The team has also added features like "privacy mode" and a "profanity filter." Researchers aim to develop a "brain-to-voice" system for natural-sounding speech.

Key takeaway

For neuroengineers and medical device developers focusing on assistive communication, this case demonstrates the critical value of long-term, independent BCI use in real-world settings. You should prioritize system automation and user-driven feature development, like privacy modes, to maximize patient autonomy and utility. This approach validates the potential for BCIs to significantly improve quality of life, even with the ongoing challenge of patient willingness for invasive procedures.

Key insights

Long-term, independent brain-computer interface use for communication and control is achievable for individuals with severe paralysis.

Principles

• BCI systems require real-world testing for patient relevance.
• Personalized decoders improve communication accuracy.
• Software updates enhance BCI functionality and independence.

Method

Doctors implant electrode arrays into the speech motor cortex. Algorithms decode neural activity into phonemes, then into words, creating personalized speech. The system is automated for independent use.

In practice

• Implement privacy and profanity filters for BCI communication.
• Automate BCI connection/disconnection for user independence.
• Expand BCI vocabulary and accuracy through continuous refinement.

Topics

• Brain-Computer Interfaces
• ALS
• Speech Decoding
• Neuroengineering
• Assistive Technology
• Neural Implants

Best for: AI Scientist, NLP Engineer, Research Scientist, Domain Expert, General Interest

Related on AIssential

• What It’s Like to Live With an Experimental Brain Implant — Brain-computer interfaces are moving from labs to medical use, offering profound benefits despite significant user commitment and technical challenges.
• Scientists reveal a tiny brain chip that streams thoughts in real time — A new single-chip BCI offers high-bandwidth, minimally invasive brain-computer communication for neurological treatment and AI interaction.
• The Overlooked Post-Stroke Crisis: How Aphasix Is Rebuilding Speech with AI — AI-driven platforms can provide structured, personalized, and measurable support for complex, underserved rehabilitation needs.
• Neural technology – Can your brain control devices or will they control your brain? — Neural technology enables direct brain-device interaction, memory manipulation, and brain activity decoding, raising significant ethical questions.
• Bionic Tech Must Prove Itself Beyond the Lab — Real-world performance and user feedback are paramount for bionic technology's long-term success and integration.
• Sensory Restoration via Brain-Computer Interfaces: A Unified 2 x 2 Framework and Convergence Roadmap — The article proposes a 2x2 framework and roadmap to unify fragmented BCI research for sensory restoration.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT Technology Review.