Objective:​ To establish a methodology to quantify the neuromuscular response to spinal manipulation, develop a comprehensive date set including factors that affect the response, and compare the responses in both healthy participants and participants with acute and chronic low back pain.

Methods:​ Surface and indwelling electromyography at eight muscle locations were recorded during lumbar side-lying manipulations in 20 asymptomatic participants, 20 acute pain participants, and 20 chronic pain participants. Onset delay detection was optimized for signal detection failures and methodological comparisons were performed using a generalized linear model. The number of muscle responses and muscle activity onset delays in relation to the manipulation contact force were compared across participant subclasses using mixed linear regressions. Effect sizes for all comparisons were calculated using Cohen’s d.

Results:​ The method of muscle activity onset delay detection that best characterized the neuromuscular response to spinal manipulation was the double-threshold method with parameters of an 8 standard deviation amplitude threshold and a 10-msec duration threshold. In healthy participants, factors such as manipulation order and location had little effect on the neuromuscular response;​ however, the responding muscle location, layer and side revealed tendencies of lower response rates, and longer muscle activity onset delays as the distance from the manipulation location increased. Symptomatic participants had less muscle responses and longer muscle activity onset delays than the asymptomatic participants. Chronic pain participants had a greater tendency for shorter muscle activity onset delays than acute pain participants.

Conclusions This study establishes a comprehensive database of both superficial lumbar and deep multifidus muscle activity and timing during spinal manipulation. The double-threshold method of muscle activity onset delay calculation is recommended over the cross-correlation method. Future studies in healthy participants focused on timing outcomes can be designed without regard for manipulation order and location within the parameters used in this study. Spinal manipulation may mediate pain through its influence on afferent activity of the muscle spindles and central nervous system. Participants in pain may experience more excitability in slower capsular reflex pathways than faster muscle spindle pathways compared to healthy participants, with the influence of the multifidi providing more pain-gating input to the central nervous system than superficial muscles. The neuromuscular response to spinal manipulation in participants in pain is dominated by the multifidus and is consistent with passive movements, as opposed to active movements that are dominated by superficial muscles.