The stability of bimanual coordinated wrist flexion and extension was examined within a paradigm of kinesthetic tracking. Eight right handed subjects were instructed to maintain one limb passive while the wrist was rotated by a computer controlled AC servo motor through amplitudes from 20 to 35 degrees either side of a "home position" set at either 30 degrees flexion or extension. Subjects were asked to track this movement with the other hand either in the same direction (simultaneous non-homologous muscle contraction: "antiphase" [AP]) or opposite direction (simultaneous homologous muscle contraction: "inphase" [IP]). Movement frequency was set for each individual at 0.2 Hz below that frequency which produced phase transitions during practice. Movement amplitude was increased from 40 degrees to 70 degrees within a trial over five 6 degree increments.

Phase transitions were observed in approximately 8 per cent of trials (41/512), almost exclusively from AP to IP. No effect of active hand or posture was revealed, indicating that either hand when constrained in flexion or extension, was equally able to maintain the AP pattern at equivalent movement rates without a transition to IP. However, there was an interaction of driven hand and posture where fewer phase transitions occurred when the right wrist was constrained in flexion, and also where the left wrist was constrained in extension, indicating these driven hand-posture conditions enabled subjects to maintain the AP pattern with greater stability regardless of the active hand posture. Measures of standard deviation of active hand posture and uniformity of relative phase revealed higher spatio-temporal stability for the flexed posture. Effects of postural constraints on spatio-temporal stability and kinesthetic tracking accuracy indicated that afference from wrist flexors and extensors respectively were not used equivalently at spinal-segmental or higher levels. These effects were explored further utilising a secondary probe reaction time task to assess demands on central capacity during IP versus AP patterns for bimanual-active versus bimanual-kinesthetic tasks. The findings are discussed in relation to the importance of kinesthetic afference and central processing demands during the performance of stable and critically unstable bimanual patterns.