Speaker : Pr. Frédéric Crevecoeur
March 29, 2021 – 1PM (English)
The field of movement neuroscience is concerned with understanding the neural mechanisms underlying purposeful movements. While moving around is likely the major if not the only purpose of nervous systems in general, there is currently no theoretical framework relating the organisation of neural networks to the time course of motor commands sent to muscles. Towards building such a framework, an important body of research has established that healthy humans use strategies corresponding to sophisticated and flexible state-feedback control models. Although these models do not tell anything about possible neural implementations, together with knowledge of the physiology they set constraints on future work that must be done to relate motor functions to neural activity. In this seminar, I will present an overview of current models of human reaching control and adaptation, and highlight the specific questions that they open for the field of AI. In particular, the distributed nature of neural correlates, the time scales at which control and adaptation unfolds, and the properties of sensory systems must be incorporated in artificial systems to explore whether artificial networks can help us understand the neural mechanisms underlying movement control.