Effects of a robot-assisted therapy with upper limb exoskeleton in subacute hemiparetic patients

Introduction: Upper limb rehabilitation after stroke functionalimpairment represents a major challenge for rehabilitation. The useof robotic exoskeletons has given preliminary efficacy demonstra-tions [1], but until now, a rigorous kinematic analysis of the effectsof these systems on the quality of movement has not been con-ducted. This pilot study aims to evaluate, through motion analysis,the effects of rehabilitation training performed with the roboticexoskeleton Armeo Power on the recovery of upper limb functionin post-stroke patients Methods: Nine sub-acute patients with post-stroke hemipare-sis were included in the study: age 57.6 ± 17.4 years, months fromevent 3.9 ± 1.5, 48.0 ± 17 Functional Independence Measure (FIM)8, Barthel Index (BI) 55.0 ± 20.8, Frenchay Arm Test (FAT) 1.7 ± 1.3,Fugl-Meyer Assessment (FMA) motor function 33.8 ± 14.2. The sub-jects were subjected to a five-week training session of 50 min ofupper limb robot therapy for 4 weeks using Armeo Power (HocomaAG), an exoskeleton with six motors and angle sensors for the sixdegrees of freedom. Clinical assessment scales (FIM, BI, FAT, FMA)and instrumental evaluations were performed at the time of inclu-sion (T0) and at the end of the training (T1). The following ArmeoPower evaluation tests have been carried out: Vertical Capture,Reaction Time and A Goal. In addition to the exoskeleton’s metric,patients were subjected to an upper limb movement analysis (RABprotocol) by means of an optoelectronic system [2] in order to studythe kinematics of the pointing motion. From the trajectory of the3rd metacarpal marker, based on 6 patient acquisitions, the follow-ing parameters [3] were calculated using Matlab (The MathWorks,Inc.): Hand path ratio (HPR), Number of Movement Units (NMU),Peak velocity (PV), Trunk Displacement (TD), Time to Peak Veloc-ity (TPV), Movement Time (MT), Normalized Jerk (NJ). Finally, theshoulder and elbow extension and shoulder abduction angles havebeen evaluated at the end of the considered task. Statistical analy-sis for all indicators was performed by Wilcoxon test (significancelevel p = 0.05).Results: At the end of the training there was a statistically signif-icant difference in the following clinical scales: FMA, BI, FAT, FIMmotor (p = 0.004). About Armeo Power evaluation tests, a statis-tically significant improvement in the Hand path ratio in VerticalCapture (p = 0.020) and reaction time (p = 0.008) was noted. Regard-ing the parameters extracted from the motion analysis, statisticallysignificant difference is noted in all parameters with the exceptionof PV and TPV: HPR, MT, NJ, TD (p = 0.008), NMU (p = 0.016). Lastly,there is a trend towards improving the compensation strategies forthe shoulder abutment, although variations in angular parametersare not statistically significant.Discussion: The overall improvement in the clinical and instru-mental parameters found provides experimental support for thequalitative results previously reported in literature: kinematicanalysis confirms the improvement of upper limb function follow-ing the use of Armeo Power in subacute post-stroke patients