Meeting Abstract
In vitro measurements reveal important muscle properties, yet their implications for animal performance are not obvious. How does neural stimulation affect jump performance? How do evolutionary transformations in skeletal anatomy influence neural control requirements? Current experimental tools are inadequate for these questions because they do not measure a muscle’s interactions with the skeleton and locomotor substrate. We used a novel technique interfacing an in vitro muscle with the ankle joint of a frog simulation containing hip, knee, ankle and TMT joints. Custom electronics transmitted in vitro force to a physics engine (MuJoCo) performing joint and ground contact dynamics in a PC. At 1 kHz, in silico muscle length was transmitted to update in vitro muscle length enabling the ‘real world’ muscle to behave as if attached to a ‘virtual reality’ skeleton whose anatomy and inertial properties were determined in software. We addressed the above questions by manipulating muscle stimulation as well as limb anatomy. The activated muscle enabled jump velocities comparable to real frogs. Advancing stimulation by 10 ms intervals (hip and knee control unchanged) caused steeper takeoff angles with speed unaffected. Additionally, manipulating the muscle’s bi-articular action at the knee limited muscle shortening, but enhanced power output by a mechanism not yet understood. Further experiments with this technique will be used to explore how dynamic and transient interactions with inertial loads and ground contact affect muscle mechanics and jumping performance.
