Novel brain-computer interface transforms visualised letters into text

A team of researchers led by the US Stanford University has developed a brain-computer interface (BCI) which allowed a patient paralised from the neck down to ‘type’ text on a digital device with a speed of 90 characters per minute and an accuracy of over 90%. The BCI combines brain implants and a neural network to decode attempted handwriting movements from neural activity in the patient’s motor cortex and then translate it to text in real time. The patient was initially asked to imagine writing 572 sentences; signals from the electrodes implanted into his brain served as input for a recurrent neural network trained to map each specific reading from the brain to the corresponding character as output. The algorithm was able to recognise the patterns produced by the brain with each imagined letter. The BCI enabled the patient to copy sentences and answer questions at a rate similar to that of someone of his age typing on a smartphone.