The aim of this article is to propose a methodology for analyzing different recovery mechanisms in subacute and chronic patients through evaluation of biomechanical parameters. Twenty-five post-stroke subjects, eight subacute and seventeen chronic, participated in the study. A 2-DoF robotic system was used for upper limb training. Two clinical scales were used for assessment. Forces and velocities at the robot's end-effector during the execution of upper limb planar reaching movements were measured. Clinical outcome measures show a significant decrease in motor impairment after the treatment both in chronic and subacute patients (MSS-SE, p<0.001; FM, p<0.05). Movement velocity increases after the robot-aided treatment in both groups. Mean values of forces exerted by subacute patients are lower than those observed in chronic patients, both at the beginning and at the end of robotic treatment, as in the latter the pathological pattern is already structured. Our results demonstrate that the monitoring of the forces exerted on the end-effector during robot-aided treatment can identify the specific motor recovery mechanisms at different stages. If the pathological pattern is not yet structured, rehabilitative interventions should be addressed toward the use of motor re-learning procedures; on the other hand, if the force analysis shows a strong pathological pattern, mechanisms of compensation should be encouraged.

The goal of this study is to propose a methodology for evaluating recovery mechanisms in subacute and chronic post-stroke patients after a robot-aided upper-limb therapy, using a set of biomechanical parameters. Fifty-six post-stroke subjects, thirteen subacute and forty-three chronic patients participated in the study. A 2 dof robotic system, implementing an "assist-as-needed" control strategy, was used. Biomechanical parameters related (i) to the speed measured at the robot's end-effector and (ii) to the movement's smoothness were computed. Outcome clinical measures show a decrease in motor impairment after the treatment both in chronic and subacute patients. All the biomechanical parameters show an improvement between admission and discharge. Our results show that the robot-aided training can contribute to reduce the motor impairment in both subacute and chronic patients and identify neurophysiological mechanisms underlying the different stages of motor recovery.