In 25F at 10A, James Haxby
Natural human experience involves a continuous stream of incoming stimuli in a rich context of prior knowledge and expectations. Traditionally, experimental psychology attempts to reduce this complexity using controlled experiments that vary a single, experimental variable and hold other, control variables constant. Human cognition, however, develops to extract information and guide behavior based on uncontrolled, naturalistic stimuli in an ecologically rich environment. In this seminar we will examine a new approach to experimental cognitive research that uses uncontrolled, naturalistic stimuli and discovers structure and meaning in the brain activity and behavioral responses they evoke using advanced computational methods from machine learning and big data analysis. We will discuss the advantages of this new approach for studying complex and ecological cognition and the limitations of the current state-of-the-art. Throughout the course we will consider future directions and challenges for extending this approach into new domains of cognition, developing richer naturalistic stimulation paradigms, and developing more powerful methods for discovering the structure of information in real world events and environments.
Approved course for the Neuroscience major/minor.
Prerequisite: Permission of instructor. Background in psychological and brain imaging research methods, computer science, and machine learning will be helpful, but students need not have background in all of these areas.