Objective ·To elucidate the causal relationship between obstructive sleep apnea (OSA) and imaging-derived phenotypes (IDPs) through two-sample Mendelian randomization (MR) studies. Methods ·The genetic data related to OSA were obtained from the genome-wide association study (GWAS) (n_{case group}=50 200, n_{control group}=401 484) in the FinnGen Biobank R11. Twenty-seven single nucleotide polymorphisms (SNPs) were screened out as instrumental variables of OSA. The GWAS of 3 935 IDPs was based on multimodal neuroimaging data from 39 691 individuals of European ancestry in the UK Biobank. Multiple MR methods, primarily utilizing inverse variance weighted (IVW) analysis, were applied, along with assessments for heterogeneity, pleiotropy, and sensitivity. Results ·MR analysis indicated that 8 IDPs were associated with OSA. For example, the genetically determined volume of caudal middle frontal gyrus in the right hemisphere was associated with an increased risk of OSA. A one‒standard-deviation increase in volume corresponded to an 11% higher risk of OSA (IVW OR=1.11, 95%CI 1.06‒1.17, P<0.001). Genetically determined reduced risk of OSA was associated with a resting-state functional connectivity characteristic (IVW β=-0.10, 95%CI -0.19‒-0.01, P=0.025), representing the functional connectivity strength between the bilateral frontal poles and the right frontal-parietal cortex. The heterogeneity test did not find significant heterogeneity among the instrumental variables. The pleiotropy test did not detect any pleiotropy. The sensitivity analysis indicated stable results. Conclusion ·Eight IDPs may have a causal relationship with the occurrence of OSA, among which one IDP shows a bidirectional causal relationship, providing potential targets for non-invasive neuromodulation interventions in OSA.