A paralyzed patient looks at a keyboard on a computer screen and types. A chip implanted in the brain detects motor signals, moves the fingers to type with speed and accuracy similar to an able-bodied person, and enters characters./Courtesy of BrainGate

U.S. scientists have developed a device that allows a patient paralyzed head to toe to type on a computer using only thoughts. Through a chip implanted in the brain, patients wrote at speeds and accuracy comparable to those of healthy people. Entering thoughts directly as text is expected to become a more familiar and easier-to-learn alternative to communication systems commonly used by paralyzed patients, such as eye-tracking systems.

A joint team from the Brigham Neuroscience Research Institute at Massachusetts General Hospital and Brown University said on the 17th in the international journal Nature Neuroscience that it developed an intracortical brain-computer interface (iBCI) typing device that allows paralyzed patients to regain communication quickly and accurately. A BCI is a technology that detects neural signals from the brain and enables the exchange of information with external devices such as computers. In other words, it is a way to translate thoughts into computer-controlled speech, text, or robotic movements.

마비 환자가 뇌에 이식한 컴퓨터 칩을 통해 컴퓨터 화면에 나타난 키보드로 일반인과 거의 비슷한 속도와 정확도로 글자를 입력하는 모습./미 매사추세츠 종합 브리검 병원

◇ Detecting motor signals in the brain, computer decodes them

Two patients, one with amyotrophic lateral sclerosis (ALS, Lou Gehrig's disease) and one with a spinal cord injury, took part in the clinical trial. The team implanted chips in the motor cortex that controls finger movements in the brains of the two paralyzed patients. The two then trained by looking at a standard keyboard with a QWERTY layout on a computer screen and entering letters.

The keyboard also shows fingers. Watching the screen, the patient thinks about the finger shape and which key to press. The chip implanted in the brain detects the brain signals generated at that moment and sends them to a computer. The computer decodes the brain signals, operates the keyboard, and inputs characters. Artificial intelligence (AI) was used to decode the brain signals. The recurrent neural network used by the team is an AI specialized for processing sequential data.

Both the patient with end-stage ALS and the patient with a spinal cord injury communicated quickly and accurately using the BCI device. One participant reached a typing speed of up to 110 characters per minute (22 words), with a word error rate of just 1.6%. That is on par with healthy people. The patients calibrated the BCI device with only 30 sentences. Notably, both used the BCI device comfortably at home, demonstrating potential for future commercialization.

Currently, paralyzed patients mainly input text into computers using eye movements. A camera tracks the gaze to determine which letter the patient is looking at on the computer screen and inputs the letter. But it takes too long to enter text on the screen, making natural communication difficult. Daniel Rubin, a Harvard Medical School professor and corresponding author of the paper, said, "BCI will be a new alternative for communication by paralyzed patients." The team said it will further increase input speed by introducing shorthand or patient-specific keyboards.

The paralyzed patient looks at a QWERTY keyboard displayed on a computer screen and thinks about where and how to move the fingers above it. A chip implanted in the motor cortex detects the signals and sends them to a computer, which decodes them to input text./Courtesy of Nature Neuroscience

◇ Clinical trials for paralyzed patients since 2004

This clinical trial was conducted as part of the BrainGate study. It is a project using BCI technology jointly developed by Brown University researchers and the bio corporations Cyberkinetics to help paralyzed patients restore bodily functions with thought alone. The BrainGate team, supported by the Ministry of Patriots and Veterans Affairs (MPVA) in the United States, has conducted clinical trials on paralyzed patients since 2004.

Leigh Hochberg, a Brown University professor and a co-author of the paper, said, "The BrainGate Consortium is validating the potential of BCI technology and greatly helping industry develop the final form of implantable medical devices for patients." Hochberg is the principal investigator of the BrainGate clinical trials and director of the Center for Neurotechnology and Neurorecovery at the Brigham Neuroscience Research Institute at Massachusetts General Hospital.

Brown University, together with Massachusetts General Hospital under Harvard Medical School, first reported results from the BrainGate clinical trials in the international journal Nature in 2006. Thanks to BCI technology, paralyzed patients could use only thoughts to control a cursor on a computer screen to open emails or operate devices such as televisions.

In 2012, the BrainGate team reported in Nature that a paralyzed patient controlled a robotic arm with BCI technology to grasp objects. Cathy Hutchinson, 58, who collapsed from a stroke and could not move below the neck or speak, succeeded in drinking coffee by moving a robotic arm as she intended thanks to BCI technology.

Fifty-eight-year-old Cathy Hutchinson, who suffered a stroke and cannot move below the neck or speak, uses BCI technology to control a robotic arm with her thoughts and drink coffee./Courtesy of BrainGate

References

Nature Neuroscience (2026), DOI: https://doi.org/10.1038/s41593-026-02218-y

Nature (2012), DOI: https://doi.org/10.1038/nature11076

Nature (2006), DOI: https://doi.org/10.1038/nature04970

※ This article has been translated by AI. Share your feedback here.