How would you calibrate a brain scan machine to happiness except by comparing it to self-evaluated happiness? You only know that certain neural pathways correspond to happiness because people report being happy while these pathways are activated. If someone had different brain circuitry (like, say, someone born with only half a brain), you wouldn't be able to use this metric except by first seeing how their brain pattern corresponded to their self-reported happiness. It seems to me that happiness simply is the perception of happiness. There is no difference between "believing you're happy" and "being happy." You can't be secretly happy or unhappy and not know it, 'cause that wouldn't constitute happiness.