Use of the boundary element method for biological morphometrics

Area of Science:

• Biomedical Engineering
• Computational Biology
• Morphometrics

Background:

• Tensorial morphometric assessments offer reference frame-independent descriptions of form differences.
• Current methods struggle with biological structures lacking landmarks and requiring domain subdivision.
• The boundary element method (BEM) can eliminate internal subdivisions.

Purpose of the Study:

• To investigate the appropriateness of a non-landmark (NL) method combined with BEM for morphometric analysis.
• To analyze the growth of the female rat neural skull from 7 to 14 postnatal days using this novel method.

Main Methods:

• Developed a non-landmark method by combining elliptical Fourier analysis (EFA) and BEM.
• Applied linear and quadratic BEM landmark analysis (10 and 5 elements).
• Constructed 500 linear BEM elements from EFA equations for NL analysis.
• Compared results with simplex triangular finite element analysis (FEA), quadrilateral FEA, and macroelement analysis.

Main Results:

• The BEM and NL methods accurately captured major growth processes in the rat skull, including cerebellar and facial vs. neural growth.
• Results showed close agreement with the biologically observed 36% area increase in other regions.
• Average differences between BEM/NL and FEA were 1.9% and 2.8%, respectively, with similar positional trends.

Conclusions:

• The combined EFA and BEM approach provides a viable non-landmark method for tensorial morphometric assessments.
• This method effectively analyzes complex biological structures without requiring numerous anatomical landmarks.
• The technique shows high accuracy and comparable results to traditional FEA methods.