Figure 2. Possible Scenarios Under the Liability Threshold Model (LTM).

The LTM proposes that, for dichotomous traits (eg, ASD), there is an underlying continuous distribution of liability for that trait in a population, typically assumed to be Gaussian. If an individual’s liability (x-axis) exceeds a specific diagnostic threshold (represented by the black vertical solid line), the individual will be considered affected. The figures assume 1 single diagnostic threshold. Panel A illustrates the conventional assumption that males and females share the same liability distribution, with the same population mean and variance. In panel B, a female protective effect is illustrated, where males and females have the same distribution for ASD liability, but the male distribution is shifted toward the singular diagnostic threshold, resulting in more males than females exceeding the diagnostic threshold. Panel C shows the same mean for both sexes, suggesting no female protective effect (but instead males have a greater variance) resulting in more males than females exceeding the diagnostic threshold. In panel D, the liability distributions between males and females differ by their means and variances. The female mean is shifted toward the diagnostic threshold, but proportionally more males cross the liability threshold than females due to a greater variance. ASD indicates autism spectrum disorder.