Noninvasive Electrical Neuroimaging of the Human Brain during Mobile Tasks including Walking and Running.

摘要

Noninvasive brain imaging during mobile activities could have far reaching scientific, clinical, and technological benefits. Electroencephalography (EEG) is the only mobile noninvasive sensing modality with sufficient temporal resolution to record brain activity on the time scale of natural motor behavior. In the past, EEG has been limited to stationary settings to prevent contamination by electromyographic and movement artifacts. I overcame this limitation by using Independent Component Analysis (ICA) to parse electrocortical processes from artifact contaminated EEG.;Chapters 2 through 4 of this dissertation demonstrate the feasibility of measuring electrocortical activity during human locomotion. In Chapter 2, subjects performed a visual target discrimination and response task while standing, walking, and running. Cognitive event-related cortical potentials during walking and running were nearly identical to those during standing. Chapter 3 provided the first intra-stride measurements of brain activity during walking. Electrocortical sources in the anterior cingulate, posterior parietal, and sensorimotor cortex exhibited significant intra-stride changes in spectral power. A substantive scientific contribution of this study is the observation that synchronous neural firing in the anterior cingulate and posterior parietal cortex, not just the sensorimotor cortex, is modulated within the stride cycle during repetitive, steady-state locomotion. A 264-channel electrode array was used in Chapters 2 and 3. By systematically reducing the number of channels used, Chapter 4 demonstrated that 35 channels were sufficient to record the most dominate electrocortical sources.;In Chapters 5 and 6, I studied healthy subjects performing isometric and isotonic lower-limb muscle contractions while seated to better understand the relationship between electrocortical dynamics and lower limb muscle activity. Isometric contractions elicited motor cortex event related desynchronization at joint torque onset and offset, while isotonic contractions elicited sustained cortical desynchronization throughout the movement. There was significant coherence between contralateral motor cortex signals and lower-limb electromyographic signals. The frequency of this coherence shifted from the beta-range for isometric contractions to the gamma-range for isotonic contractions.;This dissertation demonstrated that EEG-based brain imaging in dynamic environments is possible and expanded our understanding of cortical involvement in voluntary lower limb movement. It also provided direction for future developments of clinical neuro-monitoring, neuro-assessment, and neuro-rehabilitation technologies.