Thirty-six patients with frozen shoulders were randomized to receive either a combined central nervous system-focused (CNS) intervention with graded motor imagery (GMI) (n=18) or a peripheral-focused intervention only (i.e., manual therapy and stretching) (n=18). Outcomes of interest included active and shoulder flexion, active and passive shoulder external rotation, Shoulder Pain and Disability Index (SPADI), Patient-Specific Functional Scale (PSFS), and shoulder pain on a Visual Analog Scale (VAS) at 12 weeks, 3, and 6 months. There were no significant between-group differences at all time points for all outcomes. Within-group main effect for time point scores were significantly better in both groups after treatment with regard to SPADI scores, PSFS, VAS, and active and passive shoulder flexion and external rotation.