Brain–machine interface technology needs to be improved before it can be more widely used as a solution for limb paralysis. Prosthesis-connected electrode microarrays were implanted in the brains of two people with amyotrophic lateral sclerosis. Brain activity occurring during imagined finger movements was decoded with an algorithm and was used to control the movement of a computer cursor. The performances of the patients in this task were improved using the new algorithm compared with when they were tested using the previously available algorithm.