Background: Thoracic spinal manipulation can improve pain and function in individuals with shoulder pain; however, the mechanisms underlying these benefits remain unclear. Here, we evaluated the effects of thoracic spinal manipulation on muscle activity, as alteration in muscle activity is a key impairment for those with shoulder pain. We also evaluated the relationship between changes in muscle activity and clinical outcomes, to characterize the meaningful context of a change in neuromuscular drive. Methods: Participants with shoulder pain related to subacromial pain syndrome (n = 28) received thoracic manipulation of low amplitude high velocity thrusts to the lower, middle and upper thoracic spine. Electromyographic muscle activity (trapezius-upper, middle, lower; serratus anterior; deltoid; infraspinatus) and shoulder pain (11-point scale) was collected pre and post-manipulation during arm elevation, and normalized to a reference contraction. Clinical benefits were assessed using the Pennsylvania Shoulder Score (Penn) at baseline and 2-3 days post-intervention. Findings: A significant increase in muscle activity was observed during arm ascent (p = 0.002). Using backward stepwise regression analysis, a specific increase in the serratus anterior muscle activity during arm elevation explained improved Penn scores following post-manipulation (p < 0.05). Interpretation: Thoracic spinal manipulation immediately increases neuromuscular drive. In addition, increased serratus anterior muscle activity, a key muscle for scapular motion, is associated with short-term improvements in shoulder clinical outcomes.