Decreased diffusivity along the perivascular spaces in bipolar disorder: an MRI-based cross-sectional and Mendelian randomization study

The pathophysiology of bipolar disorder (BD) remains unclear. We investigated whether diffusion tensor imaging along perivascular spaces (DTI-ALPS), a putative MRI marker of glymphatic function, is altered in BD and whether this index relates to structural brain changes and clinical status. We recruited 108 patients with first-diagnosed and drug-naïve BD and 54 healthy controls. Participants underwent comprehensive clinical assessments (depression and mania ratings, functional disability, and neurocognition) and multimodal MRI (DTI-ALPS, free water [FW, a putative biomarker of neuroinflammation], FW-corrected mean diffusivity [MD-t], FW-corrected fractional anisotropy [FA-t], cortical thickness). We additionally assessed the causal association between DTI-ALPS and BD susceptibility using bidirectional Mendelian randomization (MR). BD patients exhibited reduced DTI-ALPS indices vs controls (P = 0.004), correlating with functional impairment (Sheehan Scale: P-corrected = 0.048) and executive dysfunction (Stroop Color and Word Test: P-corrected = 0.039). MR confirmed causal effects of lower DTI-ALPS on BD susceptibility (OR = 0.74, P = 0.01) without reverse causation. Lower DTI-ALPS indices were associated with higher FW in the forceps minor and inferior fronto-occipital fasciculus, and lower FA-t values in the forceps minor. Mediation analysis revealed that FW accumulation mediated 12.4% of the impact of alterations reflected by the DTI-ALPS index on executive performance. Together, these results provide convergent cross-sectional and genetic evidence that BD is associated with reduced DTI-ALPS, potentially reflecting impaired glymphatic fluid transport. These DTI-ALPS changes are clinically relevant, relating to disability and executive deficits in BD, and with increased FW, indicative of neuroinflammation, partly mediating this relationship.