Sleep deprivation (SD) is a prevalent problem in modern society, and one that can have serious adverse consequences for health and safety. Critically, even short periods of SD can lead to relatively large decrements in attention, which may in turn cause an individual to neglect important environmental stimuli. In this thesis, I report the results of three experiments designed to investigate the neural bases of attentional declines under conditions of sleep loss and mental fatigue. In two experiments using arterial spin labeled fMRI, a technique that enables the quantification of absolute levels of cerebral blood flow (CBF), it was found that CBF patterns in the resting brain differed significantly based on arousal levels (Study #1) and prior cognitive workload (Study #2). These findings are a departure from prior neuroimaging studies, which have typically taken neural activity during non-task periods as static and inseparable baseline. In a test of sustained attention, performance declines were observed both following SD (Study #1) and when performing the task for an extended period of time while well-rested (Study #2). These decrements were primarily mediated by hypoactivation in a fronto-parietal attentional circuit. Furthermore, resting baseline levels of cerebral blood flow in the thalamus and prefrontal cortex before the start of the task were predictive of interindividual differences in subsequent performance decline (Study #2). In Study #3, an experiment using standard BOLD fMRI, it was found that performance declines in a test of selective attention following SD were accompanied by reduced functional connectivity between top-down control areas and regions of ventral visual cortex, as well as reductions in activation to targets in object-selective areas. Taken together, these results further our understanding of the neural basis of attention under conditions when this system is taxed beyond its normal limits.