Fig. 2.

Schematic drawing for the mediation analysis. The causal estimates determined by the inverse variance–weighted (IVW) method are presented. The causal estimates toward brain volume phenotypes are from the analysis assessing head-size-normalized outcomes. The total effects were the causal estimates from two-sample MR with genetically predicted atrial fibrillation as exposure and brain volume phenotypes as outcomes. The direct effects were the causal estimates from the two-sample MR, which were adjusted for the genetic effects from ischemic stroke by multivariable MR analysis. MR Mendelian randomization