BrainGate Implant Lets Two Paralyzed People Type at 22 Words per Minute Using Attempted Finger Movements
A brain-computer interface developed by Mass General Brigham and Brown University decoded intended finger movements into QWERTY keystrokes at 110 characters per minute with a 1.6 percent word error rate, offering a faster alternative to eye-gaze systems for people with ALS and spinal cord injuries.
Overview
Researchers from Mass General Brigham and Brown University have demonstrated an implantable brain-computer interface that enabled two people with paralysis to type on a virtual QWERTY keyboard by attempting finger movements, achieving speeds of up to 110 characters per minute — or roughly 22 words per minute — with a word error rate of just 1.6 percent. The results, published in Nature Neuroscience on March 16, 2026, represent a significant step toward restoring practical, rapid communication for people who have lost the ability to speak or use their hands.
How the System Works
The neuroprosthesis is part of the BrainGate clinical trial, a long-running research program that has been testing implantable brain-computer interfaces since 2004. Microelectrode arrays placed in the motor cortex — the brain region responsible for planning and executing movement — record neural activity as participants attempt to move their fingers. A standard QWERTY keyboard layout is displayed on a screen, with each letter mapped to a specific finger and finger position: up, down, or curled. The system maps 30 distinct tokens, covering all 26 letters of the alphabet plus space and punctuation.
A recurrent neural network decodes the raw neural signals into intended keystrokes, and a predictive language model refines the output for accuracy. Critically, the system is entirely self-paced: users type as quickly or slowly as they choose, without the fixed decoding intervals that constrained earlier brain-computer interface typing systems.
Participant Results
The study involved two participants in the BrainGate trial. One participant, identified as T18, has a cervical spinal cord injury and used 384 electrodes implanted across both brain hemispheres. The other, identified as T17, has advanced amyotrophic lateral sclerosis and used 128 electrodes on one hemisphere. Neither participant was a touch typist before their injury or diagnosis, according to the Brown University announcement.
T18 achieved the peak speed of 110 characters per minute, while T17 reached 47 characters per minute. Both participants improved steadily across sessions and performed all testing from their homes rather than in a laboratory setting. The system required calibration with as few as 30 sentences before participants could begin typing.
Advantages Over Existing Alternatives
For people with severe paralysis, existing communication tools such as eye-gaze tracking systems are often slow, require frequent recalibration, and can cause fatigue. Some people with ALS eventually lose reliable eye movement as well, leaving them with no communication options. Daniel Rubin, a critical care neurologist at Mass General Brigham and senior author of the study, noted that “for many people with paralysis, when losing use of both the hands and the muscles of speech, communication can become difficult or impossible,” according to the Brown University press release.
The QWERTY-based approach leverages existing muscle memory and familiarity with keyboard layouts, which the researchers say reduces the learning curve compared to systems that use entirely novel input methods. Leigh Hochberg, director of the Center for Neurotechnology and Neurorecovery at Mass General Brigham and leader of the BrainGate trial, said the team has been “advancing and testing the feasibility and efficacy of implantable brain computer interfaces to restore communication and independence for people with paralysis” since 2004.
Broader Implications for Motor Restoration
The study’s first and corresponding author, Justin Jude, a postdoctoral researcher at Mass General Brigham, pointed to applications beyond communication. “Decoding these finger movements is also a big step toward being able to restore complex reach and grasp movements for people with upper extremity paralysis,” Jude stated in the Brown University announcement. The ability to decode fine motor intentions at the level of individual fingers could eventually feed into robotic arm control or functional electrical stimulation systems.
Limitations and Open Questions
The study has several acknowledged limitations. The sample size of two participants is small, and the two individuals had meaningfully different conditions, electrode configurations, and levels of remaining motor function. The researchers also noted that decoder accuracy degraded after several days without recalibration, a problem known as neural signal drift. Future versions of the system will need to incorporate automatic decoder updates during normal use to maintain performance over time.
The BrainGate system remains investigational and is not commercially available. Brain-computer interface technology more broadly is in a period of rapid expansion, with multiple companies and academic groups conducting clinical trials. A December 2025 STAT News overview noted that trial participation is growing from single digits to dozens of patients, and that some developers are beginning to target conditions beyond ALS and paralysis, including mental health symptoms.
What Comes Next
Rubin characterized the field as approaching a practical threshold. “BCIs are on track to become an important new alternative to what’s currently offered,” he said, according to Brown University. Jude added that there is “room to make this communication tool better, like implementing a stenography or otherwise personalized keyboard” to further increase typing speed. Whether implantable brain-computer interfaces can scale beyond small clinical trials to serve the estimated tens of thousands of people in the United States living with locked-in conditions remains an open question, but the BrainGate results demonstrate that the core decoding technology can already match practical communication thresholds.