Cognitive theorists took the a peek into that box. Festinger's elegant study on cognitive dissonance in 1959 pierced into how people think: how we resolve two conflicting ideas and how we resolve when a behavior is at odds with a belief. Functionally, via PET and fMRI scans, we can now see what areas of the brain are activated during particular tasks to narrow down what brain areas implicated in executive planning, in fear, or in anticipating a reward. These scanning techniques also allow for looking at individual differences. For example, someone with autism who views a face will not have the same area of the brain light up as someone without autism who views a face; yet that area will light up when looking at an object that holds passionate interest. (Because autism is comprised of varied constellations of symptoms, not everyone will exhibit this.)
Man sees Spot. Man pets Spot. Skinner would argue that Man has been reinforced: tail wagging or a face lick. Even watching someone else vicariously receive the reinforcement from petting can be sufficient to trigger us to reach out to pet Spot. Now it is possible to explore cognitive expectations and schemas. We can scan blood oxygenated levels in the limbic system to see peoples' level of anticipatory and consumption reward to the behavior of petting an animal (of course, squeezing Man and Spot into a scanning machine could be tricky).
Linking this to the cognitive component of CBT shows how talk therapy treatment advanced due to the cognitive neuroscience movement. As discussed in previous posts, the behavior consequence following an event is largely influenced by interpretation of the event. When our interpretation is inaccurate, problems ensue. Therapy explores life history and relational patterns and internalized messages of self that are associated with inaccurate interpretations, "cognitive errors." These errors can feed depression and anxiety and, by helping the client peek into his own black box, solutions happen.