Visual feedback is not necessary for the learning of novel dynamics

Area of Science:

• Motor control
• Human sensorimotor learning
• Robotics and biomechanics

Background:

• Humans integrate multi-modal sensory feedback (vision, proprioception) for novel sensorimotor tasks.
• Sensory feedback guides real-time movement adjustments and updates future motor commands.
• Prior research indicated adaptation to stable dynamics is possible without vision, but its necessity for internal model development remained unclear.

Purpose of the Study:

• To investigate the role of visual feedback during movement in learning stable and unstable dynamics.
• To compare learning with and without visual feedback when proprioception is available.
• To determine if visual information is essential for developing internal models of dynamics.

Main Methods:

• Examined the impact of removing visual feedback during movement on learning novel stable and unstable dynamics.
• Compared learning outcomes with and without visual feedback, with proprioception always available.
• Assessed endpoint stiffness and force post-learning to evaluate adaptation.

Main Results:

• Subjects successfully learned smooth movements in both stable and unstable dynamics, regardless of visual feedback availability.
• Adaptation to novel dynamics (endpoint stiffness and force) occurred similarly with or without visual feedback.
• Visual feedback increased movement success rate, slightly improved path straightness, and significantly reduced endpoint variability.

Conclusions:

• Visual feedback during movement is not necessary for adapting to novel stable or unstable dynamics.
• Vision's primary role appears to be fine-tuning hand trajectory corrections near movement's end.
• These findings suggest alternative sensory channels are sufficient for internalizing new motor dynamics.