Research has shown that bimanual interactions occur in tasks such as asymmetric reaching (Kelso et al, 1979), tapping rhythmically (Helmuth & Ivry, 1996) and orthogonal movements (Franz & Ramachandran, 1998). We conducted an experimental study, using a within-subjects design, to investigate coupling effects in real, imagined and phantom limb movement, as well as the impact of virtual visual feedback. In asymmetric reaching tasks we found a significant effect of real movement (F[1,20] = 6.225, p<0.05) and some effect of imagined movement, although the pattern was not clear. Phantom limb movement was found to be not significantly different from either real or imagined movement (p>0.006), because of the high level of overlap in the data. In rhythmic tapping tasks, we did not find the effect described by Helmuth and Ivry in either real or imagined movement in the controls (p>0.05). However, the amputee participants showed a clear bimanual advantage (F[1,33] = 19.531, p<0.05), indicating that the problem is not with the task paradigm. When making orthogonal movements, we found a significant bimanual effect in real movement (t[7] = -3.741, p<0.05; t[8] = -3.344, p<0.05), but not in imagined movement (p>0.05). Virtual visual feedback improved control of phantom movement, which had some impact on the coupling effects. Handedness of the participants may have had an impact on the results, and perhaps masked the interactions that occur. The impact of handedness on imagery should be carefully considered in further research. The findings suggest a complex picture of the relationships between real, imagined and phantom limb movements. They create an opening for further research into the subjective nature of imagined and phantom movements, which may grant us a better understanding of the phantom limb experience and improve our ability to treat problems that occur here.