Predicting primate tongue morphology based on geometrical skull matching. A first step towards an application on fossil hominins

Abstract

As part of a long-term research project aiming at generating a biomechanical model of a fossil human tongue from a carefully designed 3D Finite Element mesh of a living human tongue, we present a computer-based method that optimally registers 3D CT images of the head and neck of the living human into similar images of another primate. We quantitatively evaluate the method on a baboon. The method generates a geometric deformation field which is used to build up a 3D Finite Element mesh of the baboon tongue. In order to assess the method's ability to generate a realistic tongue from bony structure information alone, as would be the case for fossil humans, its performance is evaluated and compared under two conditions in which different anatomical information is available: (1) combined information from soft-tissue and bony structures; (2) information from bony structures alone. An Uncertainty Quantification method is used to evaluate the sensitivity of the transformation to two crucial parameters, namely the resolution of the transformation grid and the weight of a smoothness constraint applied to the transformation, and to determine the best possible meshes. In both conditions the baboon tongue morphology is realistically predicted, evidencing that bony structures alone provide enough relevant information to generate soft tissue.

Fig 1. 3D CT images, landmarks on bony structures and evaluation points of the upper tongue surface.

Sagittal slices for the 3D CT images of the reference human subject (a) and the baboon subject (b, c). Paired landmarks in both subjects are displayed on sagittal slices (a) and (b). The points located at the upper surface of the baboon tongue are displayed in a sagittal slice (c) and in a 3D reconstruction of the skull, the mandible and the vertebra (d).

Fig 2. FE mesh and CT image of the reference human subject.

(a) Mid-sagittal slice of the CT image with the tongue contours from the FE tongue mesh superimposed. (b) 3D reconstruction of the skull from the CT image along with FE tongue mesh inside the oral cavity. (c) Various views of the FE tongue mesh of the human reference subject.

Fig 3. Overview of proposed methodology.

Non-rigid image registration is used to estimate the tongue of the baboon subject from both the original 3D CT images and the images with bones information only. Uncertainty quantification allows the choice of registration parameters that are robust with respect to the input registration parameters.

Fig 4. Qualitative and quantitative evaluation of FE baboon tongue mesh predicted by the bone-based registration, considering the FE baboon tongue mesh predicted by the soft-tissue-based registration as the reference.

In all views, the front is on the right. Superimposed views of the meshes (pink: soft-tissue-based registration; blue: bone-based registration) on para-sagittal views at approximately 9 mm (a) and 4 mm (b) from the mid-sagittal plane, as well as on axial views at the upper mental spine (d) and at approximately 5 mm below the infradental (e). Perspective (c) and top (f) views of the 3D Spatial distribution of color-coded surface-to-surface distance between the two meshes (wireframe: soft-tissue-based registration; colored: bone-based registration).

Fig 5. Results for soft tissue registration problem.

Uncertain parameter space with accuracy (a) and Jacobian based measure J_10% (b) evaluated at 280 uncertain parameter pairs (λ₁, λ₂). Superposition of the baboon’s mid-sagittal plane with the best predicted FE tongue mesh (c) and its mid-sagittal projection (d). Top (e), lateral (f) and frontal (g) orthogonal views of the best predicted FE tongue mesh.

Fig 6. bone-based registration process.

Uncertain parameter space with accuracy (a) and Jacobian based measure J_10% (b) evaluated at 280 uncertain parameter pairs (λ₁, λ₂). Superposition of the baboon’s mid-sagittal plane with the best predicted FE tongue mesh (c) and its mid-sagittal projection (d). Top (e), lateral (f) and frontal (g) orthogonal views of the best predicted FE tongue mesh.