Neural Control of Movement Satellite Meeting

New frontiers at the intersection of cognition and motor control

The satellite is organized by:

Sam McDougle, Yale University
Dan O’Shea, Stanford University
Saurabh Vyas, Columbia University

The NCM conference has long been a venue for showcasing exemplary studies of intelligent behaviors. Many approaches have focused on how the sensorimotor system learns and generates movement. Separate studies have focused on how cognitive areas implement abstract thought and logical reasoning to guide actions over long timescales. Critically, despite the interconnectedness of these systems, current research rarely explores their interaction. How might we go about identifying and understanding the distributed circuits that implement cognitive-motor computations that convert thought into action? Perhaps the recent advent of large-scale neural measurement and manipulation technology will play a central role. However, is observation and causal manipulation sufficient, or are there still conceptual hurdles to overcome? Is this just a data science problem, or a theory problem? What does a more holistic understanding of the neural control of movement even look like?

There is a growing consensus that motor control and cognition are inherently intertwined, accompanied by a recognition that our field must focus on understanding their interaction within naturalistic behavior. However, the methods to achieve this remain unclear. We believe a central cause of this uncertainty is the absence of a unifying goal—a guiding beacon to aim for and a benchmark against which to measure progress. In essence, what will it mean to understand motor cognition? Will a deep learning model capable of goal-driven manipulation across diverse contexts suffice? Will we require a structured algorithmic understanding based on well-defined cognitive principles? Should we work backwards from neural recordings, reverse-engineering mechanisms from neural population dynamics, brain network interactions, and causal perturbations? Or should we proceed forward from tasks with well-understood solutions and clear hypotheses? Against this backdrop, the primary aim of this satellite meeting is to establish moonshot-style goals for our field in tackling motor cognition, setting forth a bold, measurable set of research objectives. Our speakers, drawn from robotics, cognitive science, and neurophysiology, will bring diverse perspectives on what progress will look like and what research programs are needed to achieve it. In panel discussions, we will focus the discussions on a “challenge question.” We will ask panelists to broadly explore what they believe should be a measurable overarching goal for the study of their topic (e.g., long-range planning), and offer perspectives on how exactly they would go about achieving that goal. These discussions should inspire innovative research collaborations, elucidate a set of open problems, and guide the emerging scholars in our community who are now asking, “What’s our field’s big question for the next decade?”

The satellite meeting will be focused on three broad areas of motor cognition that will address many of these timely and critical questions. The satellite will organize the day into three sessions: 1) long-range cognitive-motor planning, 2) skill learning and performance and its association with cognitive-motor flexibility, and 3) imitation and few-shot learning. Each session will focus on one of these three areas of motor cognition. Each will include 3–4 invited talks and a carefully designed panel discussion. To facilitate a broad discussion, each panel will be composed of experimental neurobiologists, engineers (e.g., roboticists), theoreticians, and cognitive psychologists. The satellite will have one opening and one closing keynote; one will focus on the theory and philosophy of mind while the other will focus on systems motor neuroscience.