Objective: Neuroimaging techniques have begun to elucidate the neurophysiology of bipolar disorder (BPD). Several features of BPD have hindered understanding of how mood-state changes are reflected in changes in brain physiology. Longitudinal studies have advantages in isolating state-related changes and in studying the instability, inherent in these disorders, that gives rise to pathological mood states.
Methods: To assess the state of the art in longitudinal neuroimaging studies in BPD, we conducted a literature review, searching MEDLINE for articles that included the key words bipolar disorder and magnetic resonance spectroscopy (MRS), magnetic resonance imaging (MRI), or emission tomography. The search was limited to studies with multiple subjects at two distinct and defined mood states. This search yielded eight MRS studies, four functional MRI studies, and three positron emission tomography studies.
Results: Although longitudinally designed studies allow for the isolation of biomarkers of mood state (including euthymia), the current literature is hampered by a lack of replication between studies.
Conclusions: The current body of longitudinal BPD imaging studies is heterogeneous and incomplete, and does not lend itself to the construction of an explanatory model of mood-state transitions. Drawing on extant studies, we propose a hypothetical framework for future experiments combining multimodal imaging with a longitudinal study design.