Abstract of Dr. Patricia Capsi's talk

Talk title: From Signals to Structure: Decoding Motoneuron Synergies for Physiologically Grounded Prosthetic Control

Abstract: 

Modern prosthetic hands can be mechanically sophisticated, yet their control often remains unintuitive and fails to fully exploit their capabilities. Current approaches largely focus on improving signal acquisition or combining multiple sensing modalities. In this talk, I argue that a key limitation lies instead in how motor intent is represented. I present a novel approach to human–prosthesis interfacing based on motoneuron synergies, which capture the low-dimensional structure of motor commands at a more fundamental level than muscle activity. We show that these neural synergies are more robust and span a richer control space than traditional muscle-based representations. By embedding this neural representation into a synergy-driven prosthetic hand, users can achieve continuous and natural control across a wide range of coordinated hand postures. Real-time experiments with both able-bodied participants and prosthesis users demonstrate substantial improvements in accuracy and dexterity. I will further discuss how these neural control structures adapt across different task mappings, highlighting their potential for building personalized and generalizable prosthetic controllers. Altogether, this work suggests a shift in paradigm: from decoding signals to uncovering the underlying structure of motor control—ultimately enabling more natural and predictable control for users.