Integrative approach to studying principles of natural intelligence

Research direction 1. Numerical cognition and compositionality

Research direction 2. Neural basis of relational inference in posterior parietal cortex

How does the brain make inferences based on relations between objects? To tackle this question, we have collected electrophysiology data from the posterior parietal cortex while animals performed a visual search task based on object relations in collaboration with Dr. Mehrdad Jazayeri at MIT. The large-scale neural data and its dynamical-systems analysis will shed light on how the cortical population dynamics unfold during the relational inference. We also train artificial (graph) neural networks on the relational reasoning task and open them up to examine how similar they are to biological networks. Ultimately, this project will provide an opportunity to reverse engineer natural intelligence for neuro-AI applications.

Extended data figure 9 from Lyu, Abbott & Maimon (2022) Nature

Past projects

Previous project 1. Bayesian computation through cortical latent dynamics (NHP electrophysiology)

Research questions

Previous project 2. Bayesian interval timing in human behavior & neuroimaging

Ph.D. in Sang-hun Lee’s lab at Seoul National University, March 2008 - March 2013


Previous project 3. Neuroeconomic approach to temporal discounting of monetary gain and loss

M.S. in Jaeseung Jeong’s lab at KAIST, March 2006 - February 2008