The last part of the lecture examined the extent to which brain imaging techniques could shed direct light on cognitive psychology questions concerning the content, at a given moment, of the subject's mental state. Would it be possible, from a neuroimaging examination, to infer what a person is thinking about, or what object their attention is directed towards? In theory, it's hard to see what major obstacle could stand in the way of such inferences. Brain imaging shows systematic correlations between the subject's mental state (assumed to be known) and brain activity. Since a correlation is symmetrical, it should be possible to reverse the process and infer mental state from brain activity.
Such a brain decoding strategy would have several advantages. Firstly, the ability to decode brain activation would provide the ultimate demonstration of our understanding of the brain's coding of cognitive information - if we can infer mental content from imagery, then we have understood, at least in part, in which brain regions and in what format this information is represented. We could then tackle head-on the question of how to generalize this neural code to new stimuli, and thus the degree of abstraction of cerebral representations. Thirdly, brain decoding could provide information invisible in behavior, such as a direct measure of the orientation of spatial attention. An efficient brain decoder might make it possible to dispense with any behavioral response at all. Finally, decoding could be of practical use in the field of neuro-informatics interfaces, bio-feedback, or other more controversial applications such as lie detection.
