Personal Statement

Since my foray into neurosciences, I have been fascinated with neurobiological motor control. Motor acts are the ultimate way that control how we act in the world and express our cognitive, perceptual and reflexive neural processes' outcomes. During my PhD, I focused on spinal control of movement and, during my postdoc, ventured into the cortical basis of skill learning.

Through my training, I have realized that understanding the motor basis of behavior requires an attention to basic physics, biomechanics, muscle/nerve physiology, neural networks, multiscale models, control theory principles and optimal control ideas. I have used tools of electrophysiology, optogenetics and computational analyses to explore issues in motor control. Having investigated cortical as well as spinal biological systems, it has become clear to me that the nervous system is an enormously powerful and adaptive system, and its capacity is unparalleled for flexible motor control. There are vast numbers of parallel neural processing streams in a highly evolved architecture that accomplish motor tasks, and they are also adaptable to enable new learning. How these parallel streams at cortical and spinal levels form a robust system of computational elements for seamless motor control is striking, as well as a great conceptual challenge. It is now widely agreed that several areas of the brain cooperate for motor control—the cerebral cortex, the cerebellum and the basal ganglia.

Ultimately, they exert their control on the low-level spinal cord. Kenji Doya, PhD, has proposed that different types of learning occur in these structures: cerebellum deals with supervised learning; basal ganglia is involved in reinforcement learning; and cortex covers unsupervised learning. Final execution by the low-level spinal cord is modulated by these higher central nervous system structures, but an isolated spinal cord can also execute purposeful movement. For example, a spinal frog shows precisely aimed limb wiping trajectories to remove irritants (this led early psychologists, including William James, to attribute consciousness to the spinal frog).

Similarly, a spinal cat shows some elements of coordinated locomotion (albeit, an intact cat shows voluntary locomotion, while spinal cat shows deteriorated-unbalanced bipedal coordination). These early studies indicate that there are autonomous spinal pattern generator elements that control hard-wired muscle synergies that are capable of rudimentary purposeful movements. But, higher structures also share motor representations, while offering lower-levels partial local autonomy. A cortical stroke can degrade a skilled voluntary movement partially or completely. Thus, understanding cortex's interaction with spinal cord in developing skilled movements is important. Traditionally, motor areas have been studied using reductionist approaches, but we are investigating these areas using multielectrode recording in multiple areas in nonanesthetized animals.

One of the current foci is to understand cerebello-cortical interactions post-stroke and how to modulate them for motor recovery. Over the long-term, our lab's objective is to fully understand how descending systems learn to represent motion and recruit and control the spinal cord apparatus. A related goal is to also develop neurotechnology that is geared toward developing spinal and cortical prostheses and neural interfaces that mimic or augment aspects of biological motor control in the neurologically injured.