One hundred eighty (180) patients with knee osteoarthritis (KOA) knee pain were randomized into five groups: verum acupuncture (VA, n=36), sham acupuncture (SA, n=36), celecoxib (SC, n=36), placebo (PB, n=36), and waiting list (WT, n=36). Additionally, 41 healthy controls (HCs) were included. Patients in the VA and SA groups received 10 acupuncture sessions over two weeks, while the SC group received celecoxib (200 mg daily), the PB group received a placebo capsule, and the WT group received no treatment. Resting-state functional magnetic resonance imaging (fMRI) was performed before and after treatment to assess the primary outcome of PAG rs-FC modulation. Secondary outcomes included changes in knee pain scores measured by the visual analog scale (VAS), short-form McGill Pain Questionnaire (SF-MPQ), and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). Overall, the results revealed that all treatment groups showed pain improvement. Acupuncture increased vlPAG rs-FC with brain regions involved in cognitive control, attention, and reappraisal. Compared to celecoxib and placebo, acupuncture demonstrated distinct modulation effects on PAG connectivity. These findings suggest acupuncture may exert its analgesic effects via both peripheral and central mechanisms, offering a non-pharmacological alternative for KOA pain management.