Methods for Studying the Neural Control of Natural Behaviors Using Unconstrained Marmosets

There is growing interest in the marmoset as a model species for systems neuroscience inspired by the prospect of genetic tools like those available for mice. Such tools provide opportunities to record, control and modify neural circuits and to model human disease. Phylogenetically situated between key model organisms, the marmoset also provides a unique comparative perspective on the organization of the neural components of voluntary motor control. Methods for behavioral training and neural recording to study motor control with marmosets are not well developed. Early behavioral work with marmosets suggests they may be intolerant of established techniques used with macaques, the most prevalent non-human primate model species in neuroscience research. In chapter 2, I present an approach to voluntary, automated behavioral training of marmosets validated by quantifying upper limb kinematics during foraging and demonstrating abundant voluntary engagement in experimentally useful behavior. Pioneering techniques for performing neural recordings have been developed, but there is still ample opportunity for further refinement. In chapter 3, I describe surgical techniques and a custom configuration of a wireless headstage I developed to record population responses from sensorimotor and premotor cortex. Further I demonstrate that the wireless neural recordings can be done in conjunction with quantification of upper limb kinematics and during unconstrained natural behaviors. Finally, in chapter 4, I summarize the marmoset's unconstrained behavioral repertoire and present preliminary observations of behavioral state specific structure within population responses. These observations suggest that dimensionality of population responses during natural behaviors may be higher than expected from similar results based on more highly constrained behaviors.