Heather Kosakowski

Development of Face Responses in Ventral Temporal Cortex

Faces play an important role in human social cognition. Virtually all healthy adults have cortical areas that respond more to faces than any other stimulus category. Similarly, infants have areas of cortex that respond to faces but critically lack the same selective profile seen in adults. Three theories to explain the development of face selectivity have emerged: the proto-architecture hypothesis, the subcortical template hypothesis, and the social reward hypothesis. Cortical face responses in infancy are replicable, suggesting that awake infant fMRI can provide an unprecedented look into how the infant brain supports infant cognition. Further, improvements in awake infant fMRI methods increase the resolution at which we can interrogate the infant brain to address critical questions of the development of cortical specialization. Future research using fMRI with awake infants will be able to provide data to support theories of development of face-selective cortex.

Peng Qian

Hypothesis Space Navigation in the Interpretation of Fragmentary Linguistic Input

Humans make flexible inferences from minimal, incomplete, and constrained experiences. Such inferences, however, often require navigation in a potentially unbounded hypothesis space. To study the solution of human mind to this challenging problem, we propose to investigate a simple task: filling in the blank given fragmentary linguistic input, such as 'The student put ____ desk.'. I will illustrate how this specific task setting relates to the naturalistic inference behaviors during language processing and discuss a series of intuitive examples to show that 1) various kinds of knowledge can effectively guide inferences, and 2) the characteristics of the task responses may inform us of the inference strategies in the human mind. The proposed study would help advance our understanding of robustness in language processing and contribute new insights into the algorithmic aspects of problem solving.