SpasticSim: a synthetic data generation method for upper limb spasticity modelling in neurorehabilitation.

Area of Science:

• Neurorehabilitation
• Biomechanics
• Medical Imaging

Background:

• Objective quantification of spasticity is challenging due to complex motor control.
• Heterogeneous patient data is scarce, hindering spasticity research.
• Accurate assessment is crucial for effective neurorehabilitation strategies.

Purpose of the Study:

• To generate representative synthetic human body models for spasticity research.
• To analyze muscle behavior during upper limb movements under simulated spasticity conditions.
• To create an open-source dataset for validating musculoskeletal models.

Main Methods:

• Utilized Blender and MBLab add-on to create synthetic human body models.
• Exported models to OpenSim for biomechanical simulations.
• Simulated six degrees of spasticity based on the Modified Ashworth Scale (MAS) during four upper limb movements.

Main Results:

• Generated a comprehensive, open-source dataset of synthetic patient data and movement simulations.
• Demonstrated a novel approach for creating data to study spasticity.
• Provided a foundation for testing and validating musculoskeletal models.

Conclusions:

• Synthetic data generation using Blender and OpenSim is a viable method for spasticity research.
• The open-source dataset facilitates further investigation into muscle motor disorders.
• This approach has the potential to advance neurorehabilitation by improving assessment and treatment validation.