This study investigated the differential responses of more than 100 aphasics to a “primer” program of automatic training. The subjects were exposed to random shapes on a screen, and responded by pushing buttons to match stimuli. If the response was correct, colored lights were flashed or door chimes sounded. If the response was incorrect, darkness followed. If no errors were made on certain defined trials, the program advanced the subject to jump-ahead tests, or retreated to fall-back tests if errors occurred. Therefore, it seems possible to teach severe aphasics any program which can utilize visual discrimination.