Andrea d’Avella

A modular architecture for trial-by-trial learning of redundant muscle activity patterns in novel sensorimotor tasks

This paper presents a computational model that learns redundant muscle activity patterns in novel sensorimotor tasks through trial-by-trial adaptation of muscle synergies and their recruitment coefficients.

Changes in Muscle Synergy Structure and Activation Patterns Underlie Force Field Adaptation, Retention, and Generalization

Whole-body coarticulation reflects expertise in sport climbing

The paper investigates coarticulation in sport climbing, showing that expert climbers adjust their movements based on future planned actions, while non-climbers do not.

Control limitations in the null-space of the wrist muscle system

This study investigates the ability to control muscle activation in the null-space of the wrist system for movement augmentation, finding limitations in null-space control but capability in concurrent force and co-contraction control.

Spatial and temporal muscle synergies provide a dual characterization of low-dimensional and intermittent control of upper-limb movements

The paper compares spatial and temporal muscle synergies in upper-limb movements, finding that temporal synergies require fewer components and show higher inter-subject similarity.

Independent synaptic inputs to motor neurons driving antagonist muscles

This study investigates whether co-contraction of antagonist muscles is controlled by a common synaptic input or independent inputs, finding separate neural pathways for torque and impedance modulation.

Task space exploration improves adaptation after incompatible virtual surgeries

The paper shows that task space exploration improves adaptation after incompatible virtual surgeries by enabling learning of new muscle synergies.

Mixed matrix factorization: a novel algorithm for the extraction of kinematic-muscular synergies

This paper introduces a mixed matrix factorization algorithm to extract kinematic-muscular synergies, showing improved performance over existing methods in both simulated and real data.

A Hessian-based decomposition characterizes how performance in complex motor skills depends on individual strategy and variability

This paper introduces a Hessian-based decomposition method to analyze how performance in complex motor skills depends on individual strategy and variability in high-dimensional action distributions.

An intermittent control model predicts the triphasic muscles activity during hand reaching

The paper proposes an intermittent control model that predicts triphasic muscle activity during hand reaching using unsupervised machine learning and one parameter.