Ninety patients with knee osteoarthritis were randomized to acupuncture (n=30), sham acupuncture (n=30), or waiting list (n=30). Patients received twelve 30-minute sessions over 4 weeks. The primary outcome was response (≥2-point NRS reduction at week 4). Secondary outcomes included changes in NRS, WOMAC, STAI-S, and fMRI metrics: ALFF (resting-state spontaneous activity) and seed-based functional connectivity (FC) centered on regions identified by ALFF. Outcomes were assessed at baseline and 4 weeks. Overall, the results revealed higher response with acupuncture (85.2%) versus sham (52%) and waiting (13.6%), greater NRS improvement with acupuncture versus sham, dorsolateral prefrontal cortex (DLPFC) ALFF decreases with both acupuncture and sham, stronger DLPFC–thalamus FC increases with acupuncture, and DLPFC–cerebellum FC decreases with sham. These findings suggest acupuncture’s specific analgesia engages DLPFC–thalamic circuits beyond placebo-related DLPFC modulation.