Unraveling the Local Relation Between Tissue Composition and Human Brain Mechanics Through Machine Learning

Area of Science:

• Biomechanics
• Neuroscience
• Materials Science

Background:

• Regional mechanical properties of brain tissue are critical for understanding brain injury and cellular function.
• Characterizing soft brain tissue mechanics is challenging due to its complex, heterogeneous, and time-dependent responses.
• Existing models struggle to fully capture the relationship between brain tissue microstructure and its mechanical behavior.

Purpose of the Study:

• To develop advanced computational models for characterizing regional brain tissue mechanics.
• To quantify the contribution of cellular and extracellular components to time-independent and time-dependent mechanical properties.
• To establish structure-property relationships in human brain tissue.

Main Methods:

• Combined large-strain biomechanical testing with enzyme-linked immunosorbent assays (ELISA).
• Developed an extended constitutive artificial neural network (CANN) model incorporating viscoelastic effects.
• Analyzed tissue responses across different brain regions to identify microstructural influences.

Main Results:

• The viscoelastic CANN model accurately described regional brain tissue responses.
• Extracellular matrix protein fibronectin is highly relevant for both quasi-elastic and viscoelastic behaviors.
• Basement membrane proteins influence quasi-elastic response, while cellular components are more critical for viscoelasticity.

Conclusions:

• Established novel links between brain tissue microstructure and its complex mechanical properties.
• Demonstrated the significant role of fibronectin in brain tissue mechanics.
• Findings support the development of predictive material models for simulations and biomaterials for tissue engineering.