A Methodology for Quantifying Seated Lumbar Curvatures

To understand the role seating plays in the support of posture and spinal articulation, it is necessary to study the interface between a human and the seat. However, a method to quantify lumbar curvature in commercially available unmodified seats does not currently exist. This work sought to determine if the lumbar curvature for normal ranges of seated posture could be documented by using body landmarks located on the anterior portion of the body. The development of such a methodology will allow researchers to evaluate spinal articulation of a seated subject while in standard, commercially available seats and chairs. Anterior measurements of boney landmarks were used to quantify the relative positions of the ribcage and pelvis while simultaneous posterior measurements were made of lumbar curvature. The relationship between the anterior and the posterior measures was compared. The predictive capacity of this approach was evaluated by determining linear and second-order regressions for each of the four postures across all subjects and conducting a leave-one-out cross validation. The relationships between the anterior and posterior measures were approximated by linear and second-order polynomial regressions (r 2 ¼0.829, 0.935 respectively) across all postures. The quantitative analysis showed that openness had a significant relationship with lumbar curvature, and a first-order regression was superior to a second-order regression. Average standard errors in the prediction were 5.9 � for the maximum kyphotic posture, 9.9 � for the comfortable posture, 12.8 � for the straight and tall, and 22.2 � for the maximum lordotic posture. These results show predictions of lumbar curvature are possible in seated postures by using a motion capture system and anterior measures. This method of lumbar curvature prediction shows potential for use in the assessment of seated spinal curvatures and the corresponding design of seating to accommodate those curvatures; however, additional inputs will be necessary to better predict the postures as lordosis is increased. [DOI: 10.1115/1.4005400]