Degree: Doctor of Philosophy Abstract: Background: Low back pain is a prevalent condition commonly treated with conservative care, including spinal manipulative therapy (SMT). It is well known that forces applied by SMT are transfered to spinal tissues and that these forces inititate SMT’s benefitial (or possibly harmful) health outcomes. Importantly, the distribution of these forces within spinal tissues and how it compares to forces arising from daily activities remains unclear. By identifying the distribution of these forces, it may be possible to design treatments that can specifically target, or avoid particular spinal tissues thereby making SMT intervention more effective, efficient and safe. Objective: The overall objective of this dissertation was to biomechanically investigate how loads arising from SMT application were distributed within spinal structures when using varied SMT input parameters and how these loads compared to the ones arising from passive movements. Specifically, this dissertation had four definite objectives: 1) to verify the application of the principle of superposition when testing biomechanical structures with time-dependent non-linear behavior; 2) to identify the loading characteristics of spinal tissues during SMT with different parameters of application (force magnitude and application site); 3) to investigate spinal tissues' loading characteristics when SMT is delivered using different methods of application; and 4) to describe the loads arising from manual SMT application in comparison to passive lumbar movements of the lumbar spine. Methods: To address the first objective, a stable robotic platform was used to evaluate 3D print models having time-dependent non-linear material properties. To address objectives 2-4, the following general methodology was implemented. Vertebral movement arising from SMT application with varied input parameters as well as during passive lumbar movements was quantified by optical tracking of indwelling vertebral bone pins from a cadaveric pig model. Vertebral segments were harvested en bloc and mounted in a parallel robot equipped with a 6 degree-of-freedom load cell. The parallel robot replicated the exact vertebral displacements arising from SMT applications and physiological movements while the load cell measured and recorded the loads experienced by the motion segment. By combining kinematics replication with serial dissection, loads experienced by spinal structures were measured and analysed. The four experiments that addressed objectives 2-4 applied SMT varying in force magnitude, application site and method of application, and the resulting motion segment forces and moments along and around the three Cartesian axes were then compared. Finally, motion segments' loads arising from SMT were compared with ones from passive spinal movements to provide a framework for understanding the magnitude of tissue response created by SMT. Results: The results of the first experiment suggest that even in an optimized environment with identical testing objects, the principle of superposition could not be observed: removal order and/or unique testing circumstances influence structure loading characteristics. The experiments investigating the influence of SMT input parameters on the loading characteristics of the intact specimen and spinal structures revealed that SMT input parameters of peak force magnitude and application site significantly affect SMT load distribution within spinal structures and specific spinal structures will experience unique loads as a function of the SMT input parameters of peak force magnitude and application site. Similarly, the experiment in which three different methods of SMT application were investigated revealed that the method in which SMT is applied also influenced SMT loading distribution within spinal structures and, consequently, the loading characteristics of the intact specimen and spinal structures. Finally, the comparison between the loads experienced by the intact specimen and spinal structures during SMT and passive physiological movements revealed that although loading distribution within spinal structures varied as a function of the motion applied to the spine, the forces and moments experienced during SMT were comparable to those experienced during passive physiological movements, with notable exceptions. Conclusion: Although the results reported here are specific to the order of spinal structure removal, these results provide novel evidence that it may possible to alter SMT input parameters, or use specific methods of SMT application, to specifically target particular spinal structures. Additionally, loading rate, forces and moments created by manual SMT are below previously reported injury values. The unique loading profile created by SMT may be the mechanism that confers SMT’s therapeutic effect in comparison to the loading created during daily activities. This work provides important information for clinicians about the potential impact of SMT parameters as well as a foundation for future investigations of SMT biomechanics and underlying therapeutic mechanisms.