In the present invention, voice prints are made from voiced word sounds and are used to discriminate between the command word or words sound made by the patient and the same command word or words sound made by anyone other than the patient, and to allow the task to be performed only if the patient is attempting the activation.