Data-driven modeling of heterogeneous viscoelastic biofilms.

Area of Science:

• Microbiology
• Biomechanics
• Computational Modeling

Background:

• Biofilms exhibit complex heterogeneity in structure and mechanical properties.
• Existing models often simplify biofilms as homogeneous, neglecting microscale variations.
• This simplification limits accurate prediction of biofilm behavior like deformation and detachment.

Purpose of the Study:

• To develop an image-based modeling approach for mapping biofilm mechanical heterogeneity.
• To analyze biofilm deformation and stress distribution considering spatial viscosity variations.
• To improve computational models of biofilm behavior by incorporating microscale heterogeneity.

Main Methods:

• Transformed 2D optical coherence tomography (OCT) images into pixel-scale non-Newtonian viscosity maps.
• Calibrated viscosity maps using bulk viscosity data and OCT signal intensity relationships.
• Implemented a heterogeneous Oldroyd-B constitutive model with a phase-field approach for analysis.

Main Results:

• The heterogeneous model predicted biofilm deformations more accurately than homogeneous models.
• Stress distribution analysis revealed significant differences due to viscosity heterogeneity.
• The model successfully allocated lower viscosities to low-density areas and higher viscosities to high-density areas.

Conclusions:

• The image-based, pixel-scale approach effectively captures biofilm mechanical heterogeneity.
• This novel modeling approach enhances the accuracy of biofilm deformation and stress analysis.
• Future work should refine the OCT signal-viscosity relationship and explore alternative constitutive models.