Our Research
Overview
Research in the NCC Lab focuses on the neural mechanisms underlying cognitive control, and its engagement in higher cognitive functions that underlie human intelligence. Cognitive control can be defined broadly as the ability to guide behavior in pursuit of internally represented goals and intentions, by promoting task-relevant processes over competing alternatives, often over extended periods. Historically the lab has focused on understanding the fundamental mechanisms responsible for cognitive control, and the dynamic and strategic factors that govern their deployment, providing insights into interactions among neural subsystems such as prefrontal cortex, anterior cingulate cortex, basal ganglia, and their regulation by neuromodulatory systems. As the fundamental mechanisms have come into view, work in the lab has focused increasingly on how these interact with other neural subsystems responsible for learning and memory – such as skill acquisition in posterior neocortical structures, and episodic memory in medial temporal structures including the hippocampus. In particular work has focused on understanding how these give rise to the forms of abstract representation, flexible generalization, reasoning and planning abilities that remain a unique province of human cognitive function, and are required to achieve the forms of natural intelligence of which it is capable. Toward these ends, work in the lab makes use of computational modeling, mathematical analysis, studies of human cognitive function and behavior, and neural recording techniques including fMRI, EEG and MEG. Members of the laboratory also periodically collaborate with colleagues carrying out behavioral and neurophysiological recordings in non-human species.
Although work in the laboratory is focused on normative human cognitive function, the insights gained from this work promise to provide both a deeper understanding of how disturbances of brain function can give rise to the patterns of mental function and behavior associated with neuropsychiatric conditions, as well as the design of artificial agents inspired by our growing understanding of brain function, that can more closely emulate human cognitive capabilities.
Dynamics of cognitive control
This capacity for cognitive control allows people not only to pursue goal-directed behaviors over extended periods of time, but also to rapidly and flexibly adapt behavior to novel circumstances in a fast-changing world. This flexibility requires dynamic adjustments in the allocation of control that we seek to understand through the use of well-crafted psychophysics experiments, neuroimaging studies, and neural network modeling. Where possible, we exploit the mathematical tools of control theory to construct normative models – that is, to explain how control is used to optimize behavior – and how control allocation is impacted by fundamental computational tradeoffs, such as the stability versus flexibility of representations and the use of shared (general purpose) versus separated (task-dedicated) representations in neural networks, the explore/exploit tradeoff, and how learning interacts with and shapes the balance between these tradeoffs. This work is cast within the framework of “bounded optimality,” or “resource rationality,” which assume that the human brain has evolved to optimize its computational functions subject to the constraints imposed by the tradeoffs above, and the constraints imposed by its structure and the world in which it must operate.
Agents in a Non-stationary World
We are interested in understanding the mechanisms that enable agents to operate in a complex, changing, and uncertain world. To what extent do agents explore versus exploit? How do they balance internal versus external computation? What kind of dynamics emerge when agents have multiple competing or cooperating learning systems? For example, we are interested in using modular reinforcement learning to study the idea of having “multiple selves”, in an attempt to synthesize longstanding observations of psychological conflict (i.e. in psychodynamic theories) with modern computational principles. Overall, we aim to uncover principles that explain features of human psychology and at the same time inform the design of artificial agents.
The Computational Bases of Natural Intelligence in the Human Brain
We are at a remarkable junction in the study of the human mind and brain. With increasing frequency, machines are exceeding human performance on tasks once considered to be the exclusive province of human cognition (from the powerful arithmetic capabilities of traditional symbolic computers to the efficiency and sophistication of face recognition, game playing, and natural language processing of modern deep learning neural networks). There are also other species that excel in tasks that people (and machines) can only approximate (e.g., building a spider web). However, no other type of agent – natural or artificial – exhibits the stunning range of capabilities that humans do (see figure in link below) coupled with the efficiency of learning and processing exhibited by the human brain in acquiring and executing these. That is, the human brain seems to embody both the flexibility of symbolic computation and the efficiency of artificial neural networks.
Our lab has begun to explore the mechanisms responsible for this unique profile of flexibility and efficiency of cognitive function, focusing on two critical factors: i) the ability to learn abstract, relational representations, and use these for generalization to flexibly adapt to novel settings; and ii) the ability to regulate the use of such representations – that rely on serial, control-dependent processing – when they are needed; but to strategically engage in self-reconfiguration by investing the time and effort to acquire task-dedicated representations when more efficient, automatic processing is sufficiently advantageous.