Quantification of stroke-related impairments such as abnormal shoulder and elbow joint torque coupling has been thoroughly described for both isometric torque production tasks and supported versus unsupported reaching tasks. As individuals with stroke abduct at the shoulder lifting the arm against gravity they are constrained in their ability to reach outward due to abnormal joint torque coupling with elbow flexion. The progressive effect that gravity has on functional reaching abilities remains to be described. We have made striking advances in the quantification of human movement through the implementation of robotics that will allow us to quantify reaching movements with the capacity to control for many variables such as the relative amount of gravity that the patient experiences. We expect to find progressive reductions in reaching ability as individuals with stroke are required to lift the arm against greater gravitational loads. The device may also be employed to administer quantitatively controlled rehabilitation protocols designed to increase gravitational loading of the paretic limb during functional reaching movements over a two-month intervention period. We have demonstrated the feasibility of progressive gravity reintroduction in our initial clinical intervention work involving isometric arm strengthening. Individuals with stroke were progressively trained to generate isometric joint torque patterns away from the abnormal "flexion synergy" pattern (shoulder abduction causing spontaneous elbow flexion) in an effort to demonstrate the mutability of abnormal joint torque coupling (Ellis et al., 2005). There was a significant reduction of abnormal elbow flexion coupling during maximal isometric shoulder abduction (Figure 1). Although abnormal coupling remained in part, these results demonstrated the modifiability of the underlying impairment and potential for additional recovery. We have begun to exploit this potential for further recovery by extrapolating the same methodological principles into a dynamic protocol whose functional relevance is likely to produce greater outcome. Within this protocol we progressively train individuals to reach outward against various gravitational loads essentially producing dynamic joint torque patterns outside of the abnormal coupling pattern. The current clinical intervention incorporates a double-blinded randomized controlled design that acts to progressively reintroduce gravity during reaching practice for a group of individuals with chronic stroke. A simultaneous control group protocol involves the same reaching exercises but does not require the patient to support the limb against gravity while actively reaching. The experimental group is hypothesized to have greater increases in reaching area for matched gravitational loads of the control group following the intervention. It is also hypothesized that clinical measurement scores of upper extremity function will improve to a greater degree for the experimental group. If these hypotheses are supported, progressive gravitational loading will be identified as a key factor in the rehabilitation of upper extremity discoordination. The identification of key components of rehabilitation techniques will shape the future development of neurorehabilitation research and practice allowing clinicians to access untapped potential for further functional gains. | 
