Spatial cognition is a fundamental brain function that enables us to instantly locate ourselves and others within a complexly structured three-dimensional environment. I aim at investigating yet largely unexplored brain mechanisms of such spatial cognition during natural visual experiences by uniquely combining (1) an immersive visual display for fMRI, and (2) neural response modeling using Deep Neural Networks (DNNs), which provides means for quantitative modeling of end-to-end neural computations from natural visual inputs to task-relevant outputs. The resulting models may serve as a basis for developing real-time high-throughput brain decoder to evaluate positive (e.g., immersiveness) and negative (e.g., motion sickness) effects of VR applications, and might further lead to robots and AIs with human-like spatial sense of self and others.

Additionally, I am building a novel brain data management platform to establish secure and effective sharing of a vast amount of MRI/MEG data acquired at CiNet.