An intracortical BCI that decodes attempted handwriting movements from neural activity in the motor cortex and translates it to text in real time, using a recurrent neural network decoding approach, has been developed, according to a Nature article Published on 12 May 2021.
"With this BCI, our study participant, whose hand was paralysed from spinal cord injury, achieved typing speeds of 90 characters per minute with 94.1% raw accuracy online, and greater than 99% accuracy offline with a general-purpose autocorrect," mentioned with the collaboration work of Francis R. Willett, Donald T. Avansino, Leigh R. Hochberg, Jaimie M. Henderson & Krishna V. Shenoy.
"This is an impressive study. It shows how machine-learning methods can be used to interpret neural activity at time scales fast enough to enable people with global paralysis to communicate with the world in a real-time context," said Edvard Moser, 2014 Nobel Laureate in Physiology or Medicine.
May-Britt Moser, who shared 2014 Nobel Prize in Physiology or Medicine with Edvard, agreed that "the study is a game-changer".
"Developments like these create great hopes for globally paralyzed persons who are locked in alone, without, until now, much ability to communicate anything other than very short and simple statements." While congratulated on the new study, Edvard Moser added, "It is important to recognize that BCIs are still a long way from decoding neural activity in a broader real-time communication context. The study classified neural activity from the motor cortex after repeated training on the 26 letters of the English alphabet, plus five punctuation marks. After a long training, the computer classifies letters extremely well. But with only 31 patterns to select from, the neural activity here is still very low-dimensional compared to the huge parameter space in which activity occurs during speech or thought."