How to build a dinosaur: Musculoskeletal modeling and simulation of locomotor biomechanics in extinct animals

Underpinning many aspects of paleobiological research is the concept of uniformitarianism (Hutton Reference Hutton1788 ), that certain principles and phenomena observed in the modern day have always been in action across time and space. The laws of the physical world are one such set of principles, which lend to the investigation of biological aspects that are influenced and constrained by physics, that is, biomechanics. The investigation of biomechanical phenomena in paleobiological enquiry has a long history, and at least some aspect of biomechanics has been explored in every extinct vertebrate and many invertebrate groups (Thompson Reference Thompson1917 ; Alexander Reference Alexander1989 ; Thomason Reference Thomason1995 ). However, one group in particular has received prolonged and intensive attention in this field of study: the dinosaurs, which indeed continue to lead the charge in the development and application of new biomechanical approaches to the fossil record.

Throughout the history of life on Earth, the vast majority of species to have ever existed have become extinct. Among those extinct species and lineages is to be found a staggering diversity of body forms, sizes, functions, and ecologies that have no counterpart in the modern day. Today there are no gargantuan terrestrial, aquatic, or aerial arthropods of the scale seen in the Paleozoic (Braddy et al. Reference Braddy, Poschmann and Tetlie2008 ); extant marine reptiles present only a fraction of the highly diverse phenotypes that existed in the Mesozoic (Sues Reference Sues2019 ); no modern habitat sustains the number or size of terrestrial herbivores as some evidently did in the Jurassic (Foster Reference Foster2007 ); there are no 10-tonne bipeds alive today (Hutchinson et al. Reference Hutchinson, Bates, Molnar, Allen and Makovicky2011 ); and the list goes on. Additionally, the myriad species that bridge the anatomical, physiological and ecological divide between disparate major clades today, such as “fishapods” (stem tetrapods), “mammal-like reptiles” (nonmammalian synapsids) “protobirds” (nonavian theropods), and “protowhales” (archaeocete artiodactyls) are absent from modern environments (Kemp Reference Kemp2016 ). It therefore comes as little surprise that research into the paleobiology of these enigmatic extinct species is a long-lived and still-growing field.

Methods and Results

Rather than present the techniques first and then new results obtained, here we combine the methods and findings. The reconstruction of locomotor biomechanics in a given extinct vertebrate is typically an iterative approach, wherein preliminary results obtained may signal the need for improvement in the reconstruction or modeling methodology, and therefore results can help inform methods and vice versa (see also Hicks et al. Reference Hicks, Uchida, Seth, Rajagopal and Delp2015). Such reciprocal illumination is not a case of circularity, however; so long as clear questions and standards are defined at the outset, this self-refining “total evidence” perspective can help improve the precision with which a given question is answered, maximize the robustness of results, and increase the study's transparency and repeatability. Our workflow involves the following key steps (Fig. 1):

1. 1. building digital models from three-dimensional (3D) imaging of the original fossil specimens;

2. 2. articulating digital bones together in jointed skeletons;

3. 3. delimiting joint mobility;

4. 4. calculating the 3D shape and dimensions of the whole body and its individual segments;

5. 5. reconstructing the attachments of soft tissues such as muscles or ligaments;

6. 6. quantifying the geometry of muscle paths;

7. 7. estimating aspects of muscle anatomy or physiology that influence force production; and

8. 8. using computational models for simulation and hypothesis testing of locomotor function, behavior and performance.

The ambiguity that surrounds a given unknown (and frequently unknowable) parameter, and in turn how this may affect the cascade of higher-level inferences (Witmer Reference Witmer and Thomason1995; Bates and Falkingham Reference Bates and Falkingham2018), can be more explicitly addressed through the use of this structured, hierarchical approach. It is also worth noting that, given the uncertainties associated with fossil organisms, our perspective on hypothesis testing is not one of determining “the” answer, as may sometimes seem to be the case in biomechanical studies of extant species. Instead, we seek to determine what the answer could not have been, and thereby rule out impossible and implausible solutions; what remains of the solution space after all tests have been conducted remains the realm of plausibility, subject to future testing (Blob Reference Blob2001; Gatesy et al. Reference Gatesy, Bäker and Hutchinson2009; Nyakatura et al. Reference Nyakatura, Melo, Karakasiliotis, Allen, Andikfar, Andrada, Arnold, Lauströer, Hutchinson, Fischer and Ijspeert2019).