Background: Systematic investigations are needed identifying how variability in the biomechanical characteristics of spinal manipulation affects physiological responses. Such knowledge may inform future clinical practice and research study design. Objective: To determine how contact site for high velocity, low amplitude spinal manipulation (HVLA-SM) affects sensory input to the central nervous system. Design: HVLA-SM was applied to 4 specific anatomic locations using a no-HVLA-SM control at each location randomized in an 8. ×. 8 Latin square design in an animal model. Methods: Neural activity from muscle spindles in the multifidus and longissimus muscles were recorded from L6 dorsal rootlets in 16 anesthetized cats. A posterior to anterior HVLA-SM was applied through the intact skin overlying the L6 spinous process, lamina, inferior articular process and L7 spinous process. HVLA-SMs were preceded and followed by simulated spinal movement applied to the L6 vertebra. Change in mean instantaneous discharge frequency (δMIF) was determined during the thrust and the simulated spinal movement. Results: All contact sites increased L6 muscle spindle discharge during the thrust. Contact at all L6 sites significantly increased spindle discharge more than at the L7 site when recording at L6. There were no differences between L6 contact sites. For simulated movement, the L6 contact sites but not the L7 contact site significantly decreased L6 spindle responses to a change in vertebral position but not to movement to that position. Conclusions: This animal study showed that contact site for an HVLA-SM can have a significant effect on the magnitude of sensory input arising from muscle spindles in the back.