Abstract
Although substantial progress has been made in mapping the connectivity of cortical networks responsible for conscious awareness, neuroimaging analysis of subcortical networks that modulate arousal (i.e., wakefulness) has been limited by a lack of robust segmentation procedures for ascending arousal network (AAN) nuclei in the brainstem. Automated segmentation of brainstem AAN nuclei is an essential step toward elucidating the physiology of human consciousness and the pathophysiology of disorders of consciousness. We created a probabilistic atlas of 10 AAN nuclei built on diffusion MRI scans of 5 ex vivo human brain specimens imaged at 750 μm isotropic resolution. The neuroanatomic boundaries of AAN nuclei were manually annotated with reference to 200 μm 7 Tesla MRI scans in all five specimens and nucleus-specific immunostains in two of the scanned specimens. We then developed a Bayesian segmentation algorithm that utilizes the probabilistic atlas as a generative model and automatically identifies AAN nuclei in a resolution- and contrast-adaptive manner. The segmentation method displayed high accuracy when applied to in vivo T1 MRI scans of healthy individuals and patients with traumatic brain injury, as well as high test-retest reliability across T1 and T2 MRI contrasts. Finally, we show through classification and correlation assessments that the algorithm can detect volumetric changes and differences in magnetic susceptibility within AAN nuclei in patients with Alzheimer's disease and traumatic coma, respectively. We release the probabilistic atlas and Bayesian segmentation tool to advance the study of human consciousness and its disorders. Trial Registration: ClinicalTrials.gov: NCT03504709.