Predicting Mental Rotation and Arithmetic Abilities in Adolescents Using Resting‐State Functional Connectivity

ABSTRACT

Mental rotation (MR), a core spatial ability, is linked to arithmetic performance, with meta‐analytic findings suggesting shared parietal and frontal neural correlates. While intrinsic functional connectivity has been shown to predict math ability in children, its predictive value for MR remains unclear. Baseline resting‐state functional connectivity (rsFC) of nine‐year‐olds ( n  = 1037) from the Adolescent Brain Cognitive Development study, concurrent and prospective (baseline, and 2‐ and 4‐year follow‐up) MR ability, and prospective (3‐year follow‐up) arithmetic ability were analyzed using ridge regression predictive modeling. RsFC predicted prospective arithmetic performance most robustly, whereas predictions of MR were weaker and varied across time points. Despite these differences, predictive features for MR remained relatively stable over 4 years. Notably, networks predictive of MR diverge from those predictive of arithmetic, suggesting partially distinct neural substrates, offering limited support for a shared neural processing account. Overall, rsFC appears to be a less reliable biomarker for MR than for arithmetic.