“Estimating the Parental Age Effect on Intelligence With Controlling for Confounding Effects from Genotypic Differences”, Mingrui Wang2023-06 (rare mutations, IQ)⁠:

The association between parental age at conception and children’s traits has often been studied as it may reflect germline de novo mutation accumulation and is expected to be monotonic negative. However, for IQ, the relationship has often been found to be inverted U-shaped, possibly because of confounding by parental characteristics that correlate with child-bearing age.

Here, I leverage polygenic scores (PGS) as an indirect measure of parental intelligence and examine how the effect changes as the explanatory power increase to heritability. Heritability can be estimated by calculating the phenotype variance explained by the genetic effect when the paternal-maternal ratio of the projected age effects after controlling the genetic effect matches the male-female ratio of mutation rate.

After controlling for PGS and demographic factors, I estimate a −2.0 (95% CI, −0.3 to −3.7) IQ points change in intelligence per decade rise in paternal age. After further adjustment for birth order, it declined to −0.6 (−2.6 to 1.6). Even if only the latter estimate is attributable to mutation accumulation, the result would imply a substantial contribution of de novo mutations in the variance of intelligence.

However, the association might not equal the effect of de novo mutations and further studies are needed.