Bipolar disorder (BD) is characterized by a combination of state changes that oscillate between mania and depression, combined through periods of remission, and has a clear impact on patients’ psychological and cognitive functions. Lithium treatment has been successfully tested on bipolar patients, obtaining a subjective improvement of psychological equilibrium [1]. The mechanistic explanation of why lithium is effective in mitigating bipolar symptomatology is however still unclear. Here we present the first quantitative study of the topological changes induced in the functional brain connectivity of bipolar subjects with and without lithium treatment, and compare these observations with a matched cohort of healthy subjects. The topological changes are quantified via the persistent homology of the resting state fMRI correlation matrices [2]. Using a similarity measure defined on the topological outputs [3], we are able to detect global brain changes between different groups and conditions. At the group level, patients treated with lithium display a very similar topology to that of healthy subjects, suggesting that the observed behavioural improvements might stem from its capacity to bring back the pathological connectivity to the healthy one. We also find that lithium induces an increased homogeneity across individual functional connectomes as compared with both healthy subjects and patients not treated with lithium, suggesting that lithium might induce a more rigid functional connectivity. Further, we find specific changes in how the topological features are localized in the different conditions and groups, consistent with previous work on the role of limbic and cerebellar regions in the genesis of BD[4].
