UCSF

Statement of the problem: Gait initiation and turning are fundamental human motor tasks requiring adept postural control. People with Parkinson’s Disease (PD) often exhibit postural control dysfunction during these tasks, which can be quantified using biomechanical tools. A very limited amount of research has characterized these data with the neural circuits underlying postural control in PD. In fact, it is normally not possible to record from the deep structures of the brain while performing motor tasks in humans. However, developments in deep brain stimulation (DBS) now allows the recording of local field potentials from different areas in the brain while performing motor tasks. Further research is needed to decipher the relationships between the circuitry of postural control involved in balance tasks, levodopa’s effects, and metrics of task quality because current interventions do not offer a complete resolution for postural instability in PD. This dissertation is the first body of work combine neurophysiological and biomechanical data during gait initiation and turning under varied levodopa medication states to begin to understand these phenomena for implementing effective therapies for these symptoms.Methods: Five individuals with PD exhibiting gait and balance issues were implanted with an investigational bidirectional neural interface (Summit RC+S, Medtronic Inc) connected to deep brain stimulation (DBS) electrodes at the globus pallidus and cortical paddle electrodes overlying the premotor and primary motor cortices. Subjects performed multiple gait initiation trials and 180-degree turns under “ON” and “LOW” levodopa states utilizing force plates or body worn sensors for quantifying postural control abilities. The tasks were broken into epochs to examine neural modulation across the task under differing postural control demands, as well as inputted into linear mixed models (gait initiation) or multiple linear regressions (turns) for understanding the respective influences of brain region, medication state, epoch neural data across canonical frequencies, and their relationships to postural control task metrics. Summary of findings: Much individual variation was observed among subject responses to levodopa for both dynamic neural modulation across the task and balance task quality metrics. Neural modulation across the task did not produce consistent effects on the observed task metrics. These results support theories regarding the diversity of the neural circuits underlying different balance components (i.e. they are not dopaminergic-exclusive) and PD’s variable effects. In gait initiation, low frequency power generally decreased globally in a stepwise fashion across the task from quiet standing to weight shift to stepping, where the opposite pattern was seen at higher, pro-kinetic frequencies. Coherence was also dynamically modulated across the task, with significance exhibited in influencing weight shift amplitudes and timing. Turn results were similar, with pronounced modulation between turn preparation and turn itself variably shared among subjects at beta and gamma frequencies among pallidal and cortical locations. These results offer a preliminary framework and methodology for characterizing balance in this population, suggesting its potential application in the future using adaptive and individualized neuromodulatory interventions.