Closed loop optogenetic control and thalamic state
MetadataShow full item record
Dating as far back as the eighteenth century with Luigi Galvani's seminal studies in bioelectricity, interfacing with the nervous system at fast timescales has proven invaluable for scientific investigation as well as clinical interventions in diseases such as Parkinson's and epilepsy. Traditionally, electrical stimulation has been the primary technique for neuronal control at fast timescales. Over the past fifteen years, the advent of optogenetics, a technique whereby optical excitation or inhibition of neural activity can be targeted genetically, has ushered in a new wave of experimental approaches to dissecting circuit function. To date, most optogenetic control of neural activity has been limited to open-loop stimulation. However, activity in the brain changes in a state-dependent fashion, presenting a moving target for stimulation. In contrast to open-loop stimulation, feedback control seeks to achieve target activity by updating stimulation in real-time as a function of recorded neuronal activity. In this thesis, engineering approaches to feedback control and state estimation are used to tackle the problems of controlling neuronal firing activity in vivo, with the goal of developing a set of methods that are general enough to be applied to manipulation of other types of neuronal activity. Specifically, we apply closed-loop optogenetic control (CLOC) to manipulate somatosensory thalamus, a deep brain region that serves as a central gateway for conducting sensory information to the cerebral cortex. First, we developed a design methodology for using a previously described model-free optogenetic control scheme to entrain patterns of rate modulation such as observed in the rodent somatosensory thalamus during active movement of facial whiskers. In order to ensure the optogenetic control scheme generalizes more gracefully to future multi-input/multi-output control problems, we next applied state-space model-based control and estimation techniques to the problem of manipulating thalamic firing rates. Importantly, using this approach we investigated the effectiveness of CLOC in the awake animal for the first time, as well as the response of local populations of neurons to optical stimulation rather than recording from single neurons at a time. Finally, we investigated the effect of CLOC on thalamic “state” more generally, analyze the robustness of control to a naturally-occurring disturbance (animal movement), and look at its consequences for downstream cortical activity and sensory response characteristics in the primary somatosensory pathway. As part of this analysis, a broadly-applicable state-space model based notion of thalamic state is put forth, marrying previously distinct neuroscientific and engineering notions of “state”.