Laboratory for Brain and Machine Intelligence

Laboratory for Brain and Machine Intelligence @ KAIST

Laboratory for brain and machine intelligence, KAIST

AI 2 Brain

Our research aims to understand how cognitive control is implemented in the human brain ("AI2Brain"), thereby designing brain-inspired artificial intelligent systems that show a high level of ability to perform a wide range of tasks ("Brain2AI").

In particular, we study neural computations underlying the process of a human prefrontal cortex which allocates control over behavior to multiple types of learning and inference systems. This is achieved through a combination of computational learning theory, control theory, and experimental techniques including model-based functional magnetic resonance imaging (fMRI), electroencephalography (EEG), Transcranial magnetic stimulation (TMS), and transcranial direct current stimulation (tDCS). Topics of interest include, but are not limited to, the following:


1. Reverse Engineering Meta-Control Networks of the Human Prefrontal Cortex

Our recent studies suggest that the lateral prefrontal cortex functions as a meta-controller that flexibly allocates control weights to different types of brain’s subsystems [Lee, Neuron 2014; Lee, PLOS Biol. 2015]. However, this still leaves open a fundamental question: how is it that the lateral prefrontal cortex implements an ability to take control of transitioning between subsystems? The specific aims of this research include:

(1) Inferring causal network structure of the lateral prefrontal cortex

(2) Controllability analysis of brain’s meta-control networks

(3) Stability analysis of brain’s meta-control networks

2. Neuromorphic Algorithms for Restoring Stability to Brain’s Meta-Control System

A deeper insight into mechanistic anomalies of the prefrontal meta-control permits further development of neuromorphic algorithms for restoring stability to prefrontal systems. The aims of this research are to understand how and why control processes breakdown in various psychiatric disorders, including obsessive-compulsive, eating disorders, and addiction, and to develop a neurofeedback framework to restore stability of the prefrontal meta-control system. We aim to develop:

(1) Algorithms for estimating internal states of the prefrontal meta-control system

(2) Interfaces for interaction between the prefrontal meta-control system and a computer

(3) Neurofeedback framework for treatments for mental disorders

3. Prefrontal-Hippocampal Circuits for One-shot Learning

We usually learn only gradually, incrementally forming associations between actions or events and outcomes. But every once in a while, we quickly learn to associate that stimulus with outcomes. Recently we found that a part of the lateral prefrontal cortex appears to evaluate such causal uncertainty and then activate one-shot learning when needed.  We aim to:

(1) elucidate the role of hippocampus in driving one-shot learning

(2) understand how the prefrontal-hippocampal circuit minimizes the total amount of uncertainty in the causal relationship between stimulus and outcome