Finger muscle attachments for an OpenSim upper-extremity model.

Area of Science:

• Biomechanics
• Musculoskeletal Modeling
• Human Anatomy

Background:

• Accurate musculoskeletal models are crucial for understanding human movement.
• Existing models often have discrepancies in finger muscle attachment points and moment arms compared to experimental data.
• A need exists for a validated method to refine these attachment points in computational models.

Purpose of the Study:

• To determine precise muscle attachment points for the index, middle, ring, and little fingers within an OpenSim upper-extremity model.
• To ensure these attachment points replicate experimentally measured locations and mechanical function (moment arms).
• To develop a method for scaling and translating muscle attachments while preserving mechanical function.

Main Methods:

• A two-step process was employed: first, estimating muscle function via the partial velocity method to calculate moment arms.
• Second, optimization algorithms (Simulated Annealing and Hooke-Jeeves) were used to minimize differences between experimental and modeled moment arms.
• The method involved scaling and translating muscle attachments from experimental or model environments, preserving mechanical function.

Main Results:

• The partial velocity method achieved an average variance accounted for (VAF) of 75.5% for index finger muscles.
• Optimization resulted in a mean root mean square (RMS) error of 1.5 mm, which is within one standard deviation of measured moment arms.
• Modeled muscle attachments showed small differences from experimental measurements (average < 4.9 mm), validating the technique.

Conclusions:

• The developed technique accurately estimates muscle attachment points, even for muscles lacking direct moment arm measurements.
• The resulting non-proprietary musculoskeletal model of human fingers is suitable for various applications.
• This model can enhance the understanding of complex movements like multi-touch and gestural interactions.