Modeling White Matter Microstructure to Understand Individual Differences in Intelligence

White matter microstructure is a candidate neurobiological substrate underlying individual differences in fluid intelligence, potentially through differences in neural information transfer. Yet, it remains unclear whether MRI-derived markers of white matter microstructure generalize across tracts to support latent modeling approaches. Here, we derived measurement models for markers of white matter integrity (fractional anisotropy, FA), neurite density (intra-neurite volume fraction, INVF), and myelin content (magnetization transfer ratio, MTR) across 52 tracts (HCP-1065 atlas) grouped into ten functional clusters, using data from two independent samples (N = 150, age range 20−74 years, Dortmund Vital Study, Clinicaltrials.gov: NCT05155397; N = 215, age range 18−40 years, Mainz Network Study). Confirmatory factor analyses consistently favored hierarchical bifactor models, capturing both a general factor per marker and orthogonal hemisphere-specific factors, independent of participants’ age and sample. Fluid intelligence, assessed with Raven’s Progressive Matrices 2, was significantly predicted by general factors of FA (𝛽 = 0.46, 𝑝 < .001) and MTR (𝛽 = 0.20, 𝑝 = .021) in the Dortmund sample, but not in the Mainz sample (here measured with Raven’s Advanced Progressive Matrices). These findings establish robust, anatomically informed measurement models for white matter microstructure and provide a scalable framework for investigating the biological underpinnings of cognitive abilities.