Novel Assays of Brain Networks and Applications to Neurodegeneration
Petersen, Kalen John
The human brain is comprised of spatially separable, functionally discrete networks which can be mapped in vivo using multimodal magnetic resonance imaging (MRI). MRI permits interrogation of both the neuroanatomy of healthy brain circuits and identification of neuropathological dysfunction. The dentato-rubro-thalamic tract, the major efferent cerebellar white matter pathway, has an ipsilateral branch of unknown function. Tractographic analysis demonstrates that this branch is characterized by a more medial-posterior connectivity profile in the thalamus than the classical pathway, implying partial divergence. Resting-state functional connectivity reveals bilateral correlation between activity in the cerebellar dentate and thalamus, suggesting a functional role for ipsilateral connections. This network is degenerative in neurological disorders of movement and cognition such as progressive supranuclear palsy. While cortico-cerebellar loops mediate numerous motor and non-motor processes, frontal networks are critical for behavioral self-regulation. The limbic network is characterized by feedback projections involving the ventral striatum and cortical regions including the orbitofrontal cortex and anterior cingulate gyrus, critical regulators of reward-motivated activity. The dopamine-sensitive ventral striatum is functionally altered by administration of dopamine agonists as pharmacological therapy for Parkinson’s disease. Resting-state functional connectivity is elevated in mesocorticolimbic networks centered on the ventral striatum in patients who develop aberrant impulsive and compulsive behaviors as a side effect of dopamine replacement therapy. This altered pattern is associated with enhanced reward-learning proficiency. However, portions of the limbic network have poor signal in traditional functional MRI due to susceptibility and distortion in the ventral frontal lobe. As an alternative modality, perfusion-weighted arterial spin labeling (ASL) can recapitulate brain networks generally identified with the blood oxygenation level-dependent (BOLD) signal. However, optimized image pre-processing is necessary for consistent connectivity mapping. Surround subtraction, component-based noise correction, and spatial smoothing improve matching of ASL- and BOLD-derived networks. ASL-based functional connectivity is applied to the organization of the orbitofrontal cortex, and a previously hypothesized medial-lateral division is confirmed. Medial and lateral sub-regions are shown to have different functional and structural connectivity profiles; the latter is highly connected with the salience network and thus a potential mediator of impulsivity.