Physiological tremor is an intrinsic property of the normally functioning human system which can be observed at rest or during any postural or goal-directed movement. Despite the ubiquitous nature of tremor, the issue of the tremor control has typically been ignored with such oscillations being viewed as an example of biological noise or randomness and, as such, a purposeless feature of the motor system.

A more contemporary view is that examination of tremor can provide an insight as to the organisation of the human neuromotor system. However, for almost all tremor studies, our understanding of the underlying processes generating tremor and the strategies employed to minimize limb oscillations under real world (more than one segment) situations has been limited by the practice of examining the tremor profile in a single (usually distal) limb segment. This approach is surprising given the fact that tremor is a multiple-segment problem and rarely localized to a single limb segment (Elble & Koller, 1990). It has been assumed that the inclusion of more proximal limb segments in a postural task merely produces a resultant tremor profile in the distal effector that represents the summed output from the coupled biomechanical degrees of freedom.

This paper will present results showing that the control of tremor in multiple segments is not simply the result of mechanical coupling (a common misconception) but that there is an underlying complexity that is not intuitively apparent. The coordinative solution required to minimize upper limb tremor requires not only the control of the respective degrees of freedom utilized in task performance but also of the oscillations intrinsic to each degree of freedom. Thus, the problem of tremor control is embedded within and inseparable from the degrees of freedom problem (Bernstein, 1967; Latash & Turvey, 1996).