Afferent Roles in a Motor System with Modular Organization
Vincent CK Cheung1, Andrea d'Avella2, and Emilio Bizzi3
1Division of Health Sciences and Technology, Harvard Medical School and MIT
2Santa Lucia Hospital, Rome, Italy
3Department of Brain and Cognitive Sciences, MIT
|
| One strategy the nervous system may use to coordinate the many degrees of freedom of the motor apparatus is to organize activities of muscles into discrete modules, or muscle synergies. Previous results have shown that electromyographical (EMG) data from hindlimb muscles of bullfrogs (Rana catesbeiana) during natural behaviors can be reconstructed by linearly combining 4-7 muscle synergies, each with a specific muscle amplitude structure, and a time-dependent activation coefficient. Here, we report the effects of depriving the hindlimb of feedback signals on the EMG and the muscle synergies for jumping. Bipolar electrodes were implanted into 13 hindlimb muscles, and the 7th to 9th dorsal roots, severed. EMG data were rectified, filtered, and integrated (step size = 10ms). The dimension of the data sets were then estimated by principal component analysis, and muscle synergies before and after deafferentation were extracted using non-negative matrix factorization (Lee and Seung, 1999) and independent component analysis (Hyvaerinen and Oja, 2000). In one frog, 5-synergy solutions from the intact and deafferented data sets, respectively, showed remarkable resemblence to each other, with subtle differences noted. In the synergy representing pre-extension flexion, for instance, the strong gastrocnemius (GA) component in the intact solution was missing entirely in the deafferented solution, though all other muscle components were preserved. EMG reconstructions from the extracted solutions indicate that the absence of this GA component after deafferentation corresponds to a delayed activation onset of this knee flexor in the EMG. In addition, the coefficient amplitude of an extensor synergy was attenuated after feedback deprivation, and the activation duration of another flexor synergy, increased. In another frog, preliminary analyses indicate that at least one synergy may be driven entirely by feedback signals, though further studies are needed to confirm this hypothesis. Overall, our results are consistent with the notion of muscle synergies as discrete organizing modules of the motor system. Feedback functions (1) to enhance activation of some synergies through a positive-feedback mechanism; (2) to facilitate inactivation of some synergies; and (3) to fine-tune the amplitude balance between muscles within individual synergies. Future research will focus on functional implications of this control scheme, and other possible representations of muscle synergies. Vincent CK Cheung was supported by the Chyn Duog Shiah Memorial Fellowship (awarded through the MIT Graduate Student Office). Thanks are due to Ms. Margo Cantor for technical assistance.
|
