Predicting the shape, size, and placement of adult human pubic symphyses.

Objectives: When reconstructing fossil pelves, the articulation of the pelvic bones largely relies on subjective decisions by researchers. Different positionings at the pubic symphysis can affect the overall morphology of the pelvis and the subsequent biological interpretation associated with that individual or species. This study aims to reduce this subjectivity using quantitative models to predict pubic symphysis morphology.

Methods: We collected 3D landmarks and semilandmarks on the pubic symphysis and adjacent aspects on the CT scans of 103 adults. Using geometric morphometrics we, (1) quantified pubic symphysis morphology, (2) trained simple and two-stage least-squares linear regression models to predict pubic symphysis shape, and (3) assessed the shape variation in the sample. The model with the lowest prediction error was identified as the best model. Principal components analysis was used to explore the effects of each variable on shape and hypothetical shapes were generated from the model to illustrate these effects.

Results: The best model is a two-stage least-squares model that predicts pubic symphysis size at the first stage using additive effects of sex and age, then subsequently interacts pubic symphysis size with sex and age at the second stage to predict pubic symphysis shape. Other models with low prediction errors included variables reflecting pelvic size and breadth.

Conclusion: Linear regression modeling can be used to systematically predict pubic symphysis morphology. This method can be used in addition to other techniques to improve fossil reconstructions by more accurately estimating the morphology of this region of the pelvis.