Search for a command to run...
association studies demonstrate that cognitive traits are highly polygenic, involving hundreds ofloci with small effects, yet the evolutionary dynamics governing the persistence of extremephenotypic tails remain poorly understood. We propose the Synergistic Threshold Model (STM),a quantitative framework in which cognitive fitness follows a nonlinear landscape characterizedby an adaptive optimum and a threshold-dependent penalty beyond a functional integration limit.In this model, systemizing cognition is represented as a quantitative trait shaped by manysmall-effect alleles. Assortative mating on cognitive traits inflates phenotypic variance underWright–Fisher dynamics, allowing some individuals to exceed the functional threshold even inthe absence of directional selection. Crucially, biocultural niche buffering modifies the effectivepenalty beyond this threshold, expanding the viable variance space without shifting thepopulation mean. Under these conditions, extreme phenotypes can persist as a statisticalconsequence of variance structure rather than direct selection on the extreme state itself.Forward-time Wright–Fisher simulations demonstrate that buffered subpopulations maintainstable trait means while preserving upper-tail persistence under assortative mating, whereasunbuffered populations rapidly lose extreme phenotypes. These results show thatthreshold-dependent fitness landscapes combined with demographic structure can generatelong-term persistence of extreme cognitive phenotypes without invoking single-gene causationor runaway selection.Within this framework, clinically defined autism is interpreted as a threshold-crossing outcome ofa polygenic cognitive architecture that is otherwise adaptive at moderate levels. The SynergisticThreshold Model therefore provides a population-genetic explanation for the coexistence ofadaptive cognitive variation and rare extreme phenotypes, integrating quantitative genetics,demographic structure, and niche construction into a unified evolutionary framework