In Experiment 1, eight participants generated flexion and extension movements of the wrist, in time with an auditory metronome that increased in frequency from 1.50 to 3.50 Hz (over 54 s). In separate blocks, the forearm was either prone or supine. An inverse control scheme was implemented via a torque motor system to compensate completely for the dynamics of the moving limb. It was thus possible, in separate trial blocks, to impose motion-torque dynamics that were the reverse of the original gravitational torque, or restored the original gravitational torque. On alternate trials, the participants were asked to generate flex-on-the-beat (fob), or extend-on-the-beat (eob), patterns of coordination. The principal dependent measure was the frequency at which the target pattern could no longer be maintained. When the forearm was prone, and gravity normal the fob pattern was more stable than the eob pattern. In contrast, when gravity was reversed, eob was sustained at higher frequencies than fob. When the forearm was supine, and gravity normal, eob was more stable than fob. Whereas, when gravity was reversed fob was more stable than eob. In experiment 2, eight participants performed rhythmic plantar-/dorsi-flexion movements of the foot. In separate blocks of trials, the gravitational torque acting on the moving limb segment was either normal or reversed. In each trial the frequency of a pacing metronome was increased from 1.50 to 3.25 Hz. On alternate trials, the participants were asked to generate plantar-flex-on-the-beat (pob), or dorsi-flex-on-the-beat (dob), patterns of coordination. When gravity was normal, the pob pattern was sustained at higher frequencies than dob. In contrast, when gravity was reversed, dob was more stable than pob. These results support the view that the level of efferent drive required to generate the accentuated phase of the movement cycle dictated the stability of sensori-motor coordination.
