BACKGROUND CONTEXT: Comprehending the biomechanical characteristics of the human lumbar spine is crucial for the effective management and prevention of spinal disorders. Obtaining precise material properties from patient-specific CT scans is essential for simulations to accurately mimic real-life scenarios, which is invaluable for creating effective surgical plans.

PURPOSE: The objective of this study is to leverage Finite Element Analysis (FEA) combined with Physics-Informed Neural Networks (PINNs) to investigate and predict the material properties and biomechanical responses of the human lumbar spine. This approach aims to enhance the accuracy and reliability of biomechanical simulations, providing deeper insights into spinal mechanics and their clinical applications.

STUDY DESIGN/SETTING: A FEA study with PINNs

METHODS: We developed an FEA model of the lumbar spine incorporating detailed anatomical and material properties derived from high-quality CT and MRI scans. The model includes vertebrae and intervertebral discs, segmented and meshed using advanced imaging and computational techniques. PINNs were implemented to integrate physical laws directly into the neural network training process, ensuring that the predictions of material properties adhered to the governing equations of mechanics.

RESULTS: The integration of Finite Element Analysis (FEA) and Physics-Informed Neural Networks (PINNs) enables iterative refinement of material properties and mechanical responses, achieving high accuracy in predicting properties such as Young's modulus, Poisson's ratio, bulk modulus, and shear modulus. An automated coding approach simplifies the lumbar spine model by eliminating manual tasks like segmenting and meshing, setting different contacts, and incorporating endplates and discs, significantly reducing manual work while maintaining over 94% accuracy. This method provides precise material properties, segmentation, stress fields, and surgical plans without invasive procedures, benefiting clinical practice. The integrated FEA-PINN approach effectively predicts the material properties and mechanical behavior of the lumbar spine, demonstrating reliable convergence in simulations and supporting personalized treatment plans.

CONCLUSIONS: The integration of Finite Element Analysis (FEA) and Physics-Informed Neural Networks (PINNs) in this study allows for precise and automated prediction of material properties and mechanical behaviors of the lumbar spine, significantly reducing manual input and enhancing accuracy. This approach not only ensures reliable biomechanical simulations but also supports the development of personalized treatment plans and surgical strategies, ultimately improving clinical outcomes for spinal disorders.

CLINICAL SIGNIFICANCE: The integration of Finite Element Analysis (FEA) with Physics-Informed Neural Networks (PINNs) offers significant clinical benefits by automating the segmentation and meshing of the lumbar spine. This reduces manual input while maintaining high accuracy, ensuring that biomechanical simulations reflect real-world conditions. The precise prediction of material properties enhances the reliability of biomechanical assessments, allowing for personalized treatment plans and tailored surgical strategies based on patient-specific data without invasive procedures. This approach improves surgical planning and outcomes, ultimately contributing to better patient care and recovery in spinal disorders.

The research was supported by the Boca Raton Regional Hospital (2024-2027). 

Soft Robotic Hand: Restoring Autonomy to People Suffering from Neurotrauma

Individuals who have suffered neurotrauma like a stroke or brachial plexus injury often experience reduced limb functionality. Soft robotic exoskeletons have been successful in assisting rehabilitative treatment and improving activities of daily life but restoring dexterity for tasks such as playing musical instruments has proven challenging. This research presents a soft robotic hand exoskeleton coupled with machine learning algorithms to aid in relearning how to play the piano by ‘feeling’ the difference between correct and incorrect versions of the same song.

The exoskeleton features piezoresistive sensor arrays with 16 taxels integrated into each fingertip. The hand exoskeleton was created as a single unit, with polyvinyl acid (PVA) used as a stent and later dissolved to construct the internal pressure chambers for the five individually actuated digits. Ten variations of a song were produced, one that was correct and nine containing rhythmic errors. To classify these song variations, Random Forest (RF), K-Nearest Neighbor (KNN), and Artificial Neural Network (ANN) algorithms were trained with data from the 80 taxels combined from the tactile sensors in the fingertips. Feeling the differences between correct and incorrect versions of the song was done with the exoskeleton independently and while the exoskeleton was worn by a person. Results demonstrated that the ANN algorithm had the highest classification accuracy of 97.13% ± 2.00% with the human subject and 94.60% ± 1.26% without. These findings highlight the potential of the smart exoskeleton to aid disabled individuals in relearning dexterous tasks like playing musical instruments.

The research was supported by NSF Grant No. 2205205 and NIH Grant No. R01EB025819.

Bioinspired Soft Robotic Jellyfish: Innovations in Marine Monitoring and STEM Education

Innovation

The soft robotic jellyfish stands out due to its bioinspired design, mimicking the natural propulsion mechanisms of real jellyfish. This design utilizes submersible impellor pumps connected to each hydraulic network actuator group within the jellyfish, creating an efficient and novel propulsion system. The propulsion is tested using three different actuation schemes across multiple frequencies on an inline load cell. It was found that a full stroke actuation scheme at a set of frequency produces the greatest thrust, representing a significant advancement over traditional underwater propulsion methods.

Our goal extends beyond technological innovation. We aim to leverage the appeal of the ocean's diverse and fascinating life forms to engage the students in STEM education. By teaching principles of biomimicry and soft robotics, we empower students to create their own robotic solutions that imitate subaquatic biological organisms. This project includes providing fun and engaging kitted robots for students to build, modify, and improve, and organizing public competitions for students to test their designs. This initiative aligns with the goal of developing next-generation STEM Education aquatic robotics kits that employ soft, flexible, and waterproof materials and designs, making these technologies accessible to students at various educational levels.

Feasibility/Manufacturability

The soft robotic jellyfish is designed with manufacturability in mind. The use of readily available materials and 3D printing techniques ensures that the production process is both feasible and cost-effective. The hydraulic network actuators and submersible impellor pumps are commercially available components, simplifying the manufacturing process. The adaptive bioinspired controller, developed to maintain a predetermined depth using an onboard pressure sensor, has been demonstrated to work effectively in experimental settings. This approach ensures that the jellyfish can be produced at scale without significant increases in cost or complexity.

The educational aspect involves developing enabling hardware and bioinspired soft robotic educational experiences for the students. A modular approach allows a central control canister to wirelessly communicate with pump-actuator modules that can be attached in various configurations to replicate biological organisms. Educational lesson plans will teach students the principles of biological propulsion and soft robotic actuator design and fabrication, enabling them to creatively pursue their designs.

Marketability

The soft robotic jellyfish has numerous practical applications, making it highly marketable. It can be deployed in various marine environments to continuously monitor aquatic life and ocean dynamics. Its ability to operate at depths ranging from 0 to 20 meters ensures it can cover a broad range of marine habitats. The data collected by the onboard sensors can provide valuable insights for environmental policy implementation, marine biology research, and climate change studies. The well-defined market for such a versatile and innovative tool includes governmental and non-governmental organizations focused on environmental monitoring, academic and research institutions, and industries involved in marine exploration and conservation.

Moreover, the educational kits and competitions will inspire a new generation of students to pursue careers in underwater robotics.

This research is supported by STTR Navy Phase 2 (2023-2027).