Robot-based Rehabilitation of Upper Limb Impairment in Multiple Sclerosis

In recent years, the potential of robots in the treatment of persons with motor disabilities has raised considerable interest. These devices may interact with patients by assisting or perturbing their movements, may monitor their performance and even quantify their impairment. Clinical trials involving stroke survivors clearly demonstrated that robot-therapy results in improved motor control. More recently, robot-therapy has been applied to other pathologies. A pilot study carried out by the proponents suggested that Multiple Sclerosis (MS) subjects with predominantly cerebellar symptoms may indeed benefit from robot-therapy. MS subjects with different symptoms may benefit from robot therapy as well, but the variety of symptoms and degrees of impairment that is typical of MS suggests that robots may fully show their potential if therapy is tailored on the individual subjects. In this study, the proponents aim at extending the results of the above mentioned pilot study to MS subjects with a wider variety of impairments. Based on previous studies on robot therapy with MS subjects and stroke survivors, the proponents will develop and test a robot training exercise that is specifically designed to deal with incoordination and/or muscle weakness. In a randomised controlled trial, the proponents will then assess the possible contribution of this form of robot-therapy to the rehabilitation of MS subjects. A population of clinically definite MS subjects with different degrees of upper limb impairment was randomised into 2 groups: (i) Haptic training and (ii) Sensorimotor training. Haptic training was based on a robot-assisted exercise protocol specifically designed to treat cerebellar and motor symptoms in persons with MS by counteracting, respectively, incoordination and muscle weakness. The exercise is based on a motor task - interaction with a virtual mass-spring system under the effect of a resistive load - that requires sophisticated coordination skills. Task difficulty (the time constant of the mass-spring system) and the magnitude of the resistive load (stiffness magnitude of the resistive spring) were automatically adjusted to the individual subjects' upper limb impairment. In Sensorimotor training, the task is exactly the same, but the robot generates no forces. Hand and target position are displayed on the computer screen. This group allowed quantifying the specific contribution of robot assistance to sensorimotor recovery (if any). Both groups performed 8-10 therapy sessions (1 hour/each, 2-3 times per week). Treatment outcome was assessed in terms of the Nine-Hole Peg Test (9HPT) and the ARAT scores.