Extracting within-stroke features of torque reveals how power muscles combine to act as a motor control channel


Meeting Abstract

52.2  Thursday, Jan. 5  Extracting within-stroke features of torque reveals how power muscles combine to act as a motor control channel SPONBERG, S*; FAIRHALL, AL; DANIEL, TL; Univ. of Washington; Univ. of Washington; Univ. of Washington bergs@uw.edu

Understanding locomotor control requires revealing feedback signals and their impact on body dynamics. We frequently use summary descriptions of motor outputs (e.g. mean force or velocity), which make analyses tractable, but can obscure within-cycle control consequences. However, considering the full time course of dynamics is necessary to test hypotheses of whether muscles act together as a single “control knob” (a synergy) or have independent control authority. In the hawkmoth, Manduca sexta, recent results show that timing differences between left and right downstroke power muscles (DLMs) can modulate net torque over a wingstroke. However, what specific features of torque are controlled and whether the left and right muscles independently contribute remains elusive. To test this, we recorded muscle activations and torques during > 3000 wingstrokes from six animals during left-right yawing responses to a moving visual stimulus. We used standard spike-triggered averaging and covariance analysis, as well as a new method using partial least squares (PLS) to extract the within-stroke features of torque correlated with both independent and pairwise measures of muscle activation. The PLS method produces fewer, more highly correlated features. Feature analyses extract a significant effect of muscle stroke-to-stroke timing on torque related to frequency modulation that was not detected using net torque alone. However, the left-right timing difference alone explained the visually driven torque modulation. Power muscles control yaw torque through a mechanical power differential that combines left-right muscle timing, rather than through independent left, right channels of control. These computational techniques reveal how muscles work together to control relevant features of the motor output.

the Society for
Integrative &
Comparative
Biology