I could see this successfully using face-recognition software and tied into the auditory sense. I can't remember where I read it, but I read that people on the autism spectrum respond as strongly to the emotions evoked in music as neurotypical people. An emotion decoding computer (those exist and are pretty good now) can decide whether a person's mood should be represented by a happy C chord or a sad D minor or an angry discordant sound or an stressed buzz, and the musicality would be a good non-invasive emotional injector/empathizator.