TY - JOUR T1 - Gravitational signals underlie tonic muscle activity during goal-directed reaching JF - bioRxiv DO - 10.1101/138263 SP - 138263 AU - Erienne Olesh AU - Bradley Pollard AU - Valeriya Gritsenko Y1 - 2017/01/01 UR - http://biorxiv.org/content/early/2017/05/15/138263.abstract N2 - Human reaching movements require complex muscle activations to produce the forces necessary to move the limb in a controlled manner. How the complex kinetic properties of the limb and gravity contribute to the generation of the muscle activation pattern by the central nervous system (CNS) is a long-standing question in neuroscience. One common theory is that the CNS reduces the redundancies and complexities of the musculoskeletal system using motor primitives. These primitives are often obtained using decomposition methods based on shared variance across multiple signals. A critique of this technique is that the dependencies that exist due to the causal relationship between muscle activations and the resulting movement are difficult to disambiguate from neural primitives inherent in control signals. In the present study addressed this critique by examining the relationships between motor primitives extracted from muscle activity, muscle torques, and other motion signals. We hypothesized that the primitives obtained from muscle activity are more similar to kinetic primitives obtained from joint torques, than kinematic primitives obtained from joint angles and angular velocity signals. Eight healthy subjects pointed in virtual reality to visual targets arranged to create a standard center-out reaching task in three dimensions. Muscle activity and motion capture data were synchronously collected during the movements. Non-negative matrix factorization was then applied to muscle activity, muscle torques, and other motion signals (joint angles, angular velocities, gravitational torques, and other inertial torques) separately to reduce the dimensionality of data. Results show that the activation profiles of all NMF components were organized sequentially and correlated highly. The scaling of NMF components obtained from EMG and kinetic and kinematic signals correlated across multiple signal types. We found closer correspondence between NMF components obtained from EMG and gravitational torques, than those obtained from other torque signals or kinematic signals. Altogether, these results reject our hypothesis, suggesting that motor primitives do not consist of signals of a single modality. Our results also identify the kinetic signals for gravity compensation as the potential contributor to neural motor primitives that may be responsible for controlled transitions between arm postures during movement. ER -