The Neurodynamic Basis of Real World Face Perception

 Understanding how our brains process information while we interact with the real world is a central objective for neuroscience. However, most important neuroscientific discoveries have come from studying brain activity that was recorded while people performed tightly controlled laboratory experiments, which leaves us with open questions about how those findings relate to the brain in the real world. Recent advances in technology have made it possible to record the natural environment, behavior, and brain activity simultaneously and at scale, making it possible to study the brain in the real world. However, realizing the potential of these advances for scientific discovery requires confronting two intertwined questions: Can we even model the uncontrolled variability that arises in the real world? And if we can, then can we learn anything about the brain by doing so? This thesis attempts to answer these questions in the context of face perception during natural social interactions. It introduces methods that address the engineering and analytical challenges necessary to harness large datasets and transform the uncontrolled variability in real world behavior from a challenge into an asset that enables scientific discovery