Abstract
Low back pain (LBP) is a common cause of activity limitation in individuals that can result in socioeconomic costs up to $200 billion per year. Most cases of LBP lack a known underlying pathology. The L4/L5 motion segment is the most impaired lumbar segment, likely due to high load-bearing function. The ability to model L4/L5 compressive loading from surface electromyography (sEMG) data during dynamic activity may add to the understanding of LBP. Eight volunteers with no history of LBP participated in this study. Muscle activity of the erector spinae, rectus abdominus, and external obliques were recorded by a wireless EMG system (Trigno, Delsys, Natick, MA, USA) during a straight-leg stoop-to-stand task. L4/L5 compressive loading was estimated using a subject-specific sEMG model and validated by comparison with an AnyBody model and publicly available data from OrthoLoad. A specific trendline showed a significant decrease in percent error of estimated force for all muscles. Significantly lower impulse values were estimated by the AnyBody model than the sEMG subject-specific model (p = 0.007). Although our sEMG model was subject to high variability, loading values largely remained within those reported in the literature. Significant variation was found comparing the sEMG model with the AnyBody model, which may validate continued development and testing of personalized measurements of L4/L5 loading.