Neuroimaging is a powerful non-invasive method for studying brain alterations in bipolar disorder (BD). To date, most neuroimaging studies of BD include smaller cross-sectional samples reporting case versus control comparisons, revealing small to moderate effect sizes. In this narrative review, we discuss the current state of MRI-based, structural imaging studies, which inform our understanding of altered brain trajectories in BD across the lifespan. Alternative methodologies such as those that model patient deviations from age-specific norms are discussed, which may help derive new markers of BD pathophysiology. We discuss evidence from neuroimaging genetics and transcriptomics studies, which attempt to bridge the gap between macro-scale brain variations and underlying micro-scale neurodevelopmental mechanisms. We conclude with a look toward the future and how ambitious investments in longitudinal, deeply phenotyped, population-based cohorts can improve modeling of complex clinical factors and provide more clinically-actionable brain markers for BD.