From Arm to Leg: A Shared Primitive Framework for Solving Inverse Kinematics in Human Movement

Abstract

In motor control, the idea of a modular control architecture has been repeatedly discussed. This framework has been widely applied to kinematic and kinetic data, primarily to examine how the central nervous system (CNS) controls multiple elements simultaneously and how it handles redundant or excessive degrees of freedom. However, the inverse kinematics problem, traditionally defined as the mapping from a desired task-space outcome to the corresponding joint-space solution, has usually been treated separately from this framework. Recent work from our laboratory suggests that modular control may directly address all these problems by relying on a shared low-dimensional primitive space common to internal and external movement representations, as supported by recent studies of human hand-drawing kinematics. In the present study, we extended this framework from upper-limb drawing to lower-limb gait kinematics. We also used it to explore how different CNS pathologies may alter the organization of movement within this shared space. Overall, this study investigates whether walking, like other forms of movement, can be understood as arising from a small number of underlying primitives that are flexibly combined across coordinate systems.