Artificial neural network supported design of a lattice-based artificial spinal disc for restoring patient-specific anisotropic behaviors

To tackle the challenge in artificial spinal disc (ASD) design of restoring mechanics a natural disc, this study proposes an innovative lattice-based ASD for reproducing patient-specific anisotropic rotational response, inspired by freedom provided lattice structures. Motivated great potential machine learning to improve computational processes, method is proposed computationally efficient topology optimization using neural networks (ANNs) and subsequent member sizing automating ASDs. The results reported show good match between optimized ASDs’ six stiffnesses with those both L2-L3 L4-L5 human lumbar discs. Additionally, fast convergence rate verifies application ANNs strategy reduce space formulating problem as optimizing unit-cell distribution predefined grid. Therefore, demonstrates that able reproduce response machine-learning-based improves efficiency automated process produce personalized designs.