Professor Tal Yarkoni | What can large-scale meta-analytic modeling of fMRI data tell us about the structure of human cognition?

The search for the structures underlying human cognition, and their corresponding neural substrates, has preoccupied many psychologists and neuroscientists in recent decades. How should scientists best carve up the many dimensions of cognition, emotion, and action into distinct functions or faculties? Do the terms "working memory" and "executive control" reflect the same or different underlying processes? Are there such things as "basic" emotions, and if so, how many? Such questions are difficult to answer but important to ask. Unfortunately, none of them will be directly addressed in this talk. Instead, I will focus on a number of important methodological and conceptual issues surrounding our collective efforts to study the structure of the human mind via brain imaging methods. These include low sensitivity and specificity, poor construct validity, and a lack of isomorphism between constructs at different levels of description. Drawing on a series of recent studies, I demonstrate how large-scale meta-analyses of thousands of published fMRI studies can help us overcome many, but perhaps not all, of these issues. I conclude with a speculative discussion of the short-term and long-term prospects for a fully realized cognitive neuroscience of the human mind.