Optogenetic functional magnetic resonance imaging investigation of brain network dynamics

Abstract

Understanding how the brain rapidly processes neural information through large-scale network dynamics is fundamental to neuroscience. Resting-state functional magnetic resonance imaging (rsfMRI) has emerged as a robust, non-invasive imaging technique for mapping brain-wide functional networks, with significant correlations to sensory, behavioral, and cognitive task performance. Previous studies have primarily used sensory and cognitive tasks to investigate the role of rsfMRI networks in neural information processing. However, these external stimuli often evoke complex, long-lasting neural activities that intermingle and cascade with information flows from multiple parallel pathways. At present, whether rsfMRI networks can accommodate fast, sensitive neural information processing, including the underlying mechanisms, remains unclear. The primary objective of this doctoral work is to develop and implement the integration of cutting-edge neuromodulation, simultaneous rsfMRI with electrophysiology recordings, and dynamic network analysis for in vivo examination of the dynamic nature of brain-wide rsfMRI networks and their contributions to prompt neural information flow and processing. In this study, we reveal brain-wide rsfMRI network dynamics (i.e., states and their transitions) that enable rapid information processing through flexible activity-triggered network reconfiguration in rodent models. Using data-driven dynamic network analysis, we uncovered transient rsfMRI network states and their transitions with specific evolving spatiotemporal patterns. By applying single-pulse optogenetic stimulation to the somatosensory thalamus, we bypassed parallel sensory pathways typically engaged by external stimuli and avoided the complex cascades and excessive synchronization of neural populations caused by prolonged or repetitive stimulation. We demonstrated that this approach exhibited hierarchical somatosensory neural activity propagation and minimized disruption of ongoing spontaneous neural activities, allowing us to uniquely study the mechanisms behind rsfMRI network dynamics. Single-pulse stimulation rapidly altered the probability profiles of states and occurrence rates of their transitions, with alterations occurring within approximately 2 s and normalizing around 10 s. Specifically, following activation of a single somatosensory thalamic activity input, rsfMRI network states and transitions associated with internally-oriented information processing were rapidly suppressed and diverted through the basal forebrain and hypothalamus to strengthen externally-oriented (i.e., sensory) processes. Simultaneous rsfMRI and electrophysiology measurements in the somatosensory and cingulate cortices demonstrated that intra-regional wide-band and inter-regional narrow-band (i.e., slow oscillations) neural activity synchronizations directly drove rsfMRI network dynamics and their reconfiguration to support the rapid processing of the single thalamic input. Taken together, our study uncovers the fast interactions between spatiotemporal neural activity synchronizations and brain-wide rsfMRI networks in processing a single neural input. They demonstrate the intrinsic capability of rsfMRI networks to rapidly process neural activity, allowing the brain to respond to ever-changing demands.