Task-specific differentiation of patellofemoral pain in women using a neural network analysis of joint angle subsets

Background

Individuals with patellofemoral pain (PFP) present mechanical disorders during functional tests and activities. Studies often evaluate this population to understand the relationship between mechanical factors and pain/malalignment.

This study aimed to identify how different groupings of kinematic variables can differentiate asymptomatic women from women with PFP in various functional tasks.

Methods

Thirty-five women with PFP and thirty-five women without PFP underwent three-dimensional kinematic evaluation of seven functional tasks. All were physically active and aged between 18 and 30 years old. Seven kinematic variables were selected and grouped into four sets for each task and group (PFP and control): Proximal, Local, Distal and Malalignment. The kinematic sets differentiate the groups using the neural network-based Movement Deviation Profile (MDP). To compare the magnitude of the MDP results of the variable sets, Z-scores were used, expressing the MDP results in units of standard deviation of the control group after mean correction.

Results

The most discriminative sets varied by task. The Distal set best separated the groups during landing, and the Proximal+Local set during propulsion. The Proximal set was most discriminative for stair ascent, while the Proximal+Distal set performed best for stair descent. Malalignment variables showed greater differences in lateral step down, forward step down, and walking.

Conclusion

Kinematic variable sets, alone or in combination, can distinguish women with PFP from asymptomatic individuals in a task-specific manner. Women with PFP appear to adopt adaptable movement strategies depending on the task. These findings may improve clinical assessment and guide treatment.