In this study, the effect of dynamic stabilizers on the patellofemoral (PF) joint was investigated in normal volunteers (group I) and in patients with patellar pain (group II) or instability (group III) by using computed tomography (CT) analysis and integrated electromyography (iEMG) of the quadriceps muscle. Nine subjects (16 knees) from group I, 10 patients (12 knees) from group II and 8 patients (12 knees) from group III were included in the study. CT scans of the PF joint with quadriceps contracted (QC) and uncontracted (QU) and iEMG of vastus medialis obliquus (VMO), vastus lateralis (VL) and rectus femoris (RF) were obtained with the aid of a specially designed jig at 0 degree, 15 degrees, 30 degrees and 45 degrees of knee flexion. The same muscle contraction pattern simulating closed kinetic chain exercise was used for both CT and iEMG. The difference between the congruence angles (CA) and tilt angles (PTA) in QC and QU positions and VMO:VL ratio from the iEMG were calculated separately for each flexion angle. CA was increased in all groups with quadriceps contraction at 0 degree and 15 degrees of flexion. PTA was decreased in group I and increased in groups II and III with quadriceps contraction at the same flexion angles. This difference was statistically significant in group III at 0 degree and 15 degrees of flexion. Quadriceps contraction did not affect the patellar position significantly even in the instability group at 45 degrees of flexion. In all flexion angles the balanced VMO:VL activity ratio was observed only in group I. In the other groups, VL activity was higher than VMO activity except at 45 degrees of flexion. These findings do not support the hypothesis of dominant centralizing effect of VMO on the patella in extension, but the effect of the VMO may be more clearly demonstrated by measuring PTA in both QC and QU positions.