Lithium effects in the frontolimbic circuitry: a systematic review of neuroimaging findings in bipolar disorder.

Abstract

<h4>Introduction</h4>Lithium is a cornerstone treatment for bipolar disorder, but the biological mechanisms underlying its therapeutic effects remain incompletely defined. We adopt a network-level, cross-modal perspective focused on frontolimbic circuitry and explicitly examine links between neuroimaging findings and clinical lithium response.<h4>Methods</h4>Following PRISMA guidelines, we searched PubMed and Embase to September 3, 2025. Studies of humans were eligible when they compared lithium-treated bipolar subjects with non-lithium bipolar subjects, other treatment, healthy controls, or within-subject baselines. Risk of bias was assessed with RoB 2, ROBINS-I, NOS/NOS-CS, and BrainNet Europe criteria, as appropriate.<h4>Results</h4>We included 115 studies (randomized, n = 6; non-randomized, n = 25; observational cohort, n = 2; cross-sectional, n = 78; post-mortem, n = 4). Across modalities, lithium treatment was most consistently associated with larger volumes in the hippocampus, amygdala, anterior cingulate cortex (ACC), and prefrontal cortex (PFC), and with higher white-matter integrity within frontolimbic circuitry. Spectroscopy and functional findings were more heterogeneous in cross-sectional designs. Only a minority of studies directly related imaging to clinical response, converging on increased amygdala-PFC/ACC connectivity, hippocampal integrity, and hippocampal, amygdalar, PFC, and ACC volumes. Overall risk of bias was substantial, yielding low-to-moderate certainty of evidence.<h4>Discussion</h4>Findings support the hypothesis that lithium's therapeutic effects may be mediated by modification of the frontolimbic circuitry, while underscoring the need for preregistration, rigorous confounding control, standardized imaging pipelines, and longer follow up to capture duration-dependent effects. Ultra-high field lithium MR imaging and imaging transcriptomics may enable the development of connectomic biomarkers predictive of clinical response.