Researchers from the Harvard Brigham and Women's Hospital Neuroscience Institute and Brown University have created a brain-computer interface that allows text to be typed using only the power of thought.
The technology involves implanting sensors in the brain's motor cortex. These sensors read signals generated when the patient imagines their fingers moving while typing on a virtual QWERTY keyboard . Artificial intelligence algorithms then convert the "noise" emitted by the neurons into text.
The system matches each "micro-movement" (bending or straightening a specific finger) with a specific letter. Then, artificial intelligence algorithms process this data, converting the "noise" of neurons into meaningful words and sentences.
Commenting on this, the project's lead researcher and neurologist, Daniel Rubin, said: "Current alternative communication systems often disappoint patients due to their slowness and errors. The artificial neural prostheses we are developing aim to be a complete and rapid replacement."
Two patients participated in the trials: one with amyotrophic lateral sclerosis (ALS) and the other with a spinal cord injury . Only 30 sentences were used for calibration . One participant achieved a typing speed of 110 characters per minute (approximately 22 words), with an error rate of 1.6%, performance comparable to that of a healthy individual. The device was operated not only in the laboratory but also at home.
Neurologist Daniel Rubin stated, "Our artificial nerve limbs are intended to be a rapid replacement."
According to one of the study's authors, Justin Judd, this is a step toward restoring the complex functions of manipulating objects. The researchers plan to develop specialized keyboards and neural shorthand technologies in the future .
