Somatosensory system integrity explains differences in treatment response after stroke

摘要

<h3>Objective</h3> To test the hypothesis that, in the context of robotic therapy designed to enhance proprioceptive feedback via a Hebbian model, integrity of both somatosensory and motor systems would be important in understanding interparticipant differences in treatment-related motor gains. <h3>Methods</h3> In 30 patients with chronic stroke, behavioral performance, neural injury, and neural function were quantified for somatosensory and motor systems. Patients then received a 3-week robot-based therapy targeting finger movements with enhanced proprioceptive feedback. <h3>Results</h3> Hand function improved after treatment (Box and Blocks score increase of 2.8 blocks, <i>p</i> = 0.001) but with substantial variability: 9 patients showed improvement exceeding the minimal clinically important difference (6 blocks), while 8 patients (all of whom had &gt;2-SD greater proprioception deficit compared to 25 healthy controls) showed no improvement. In terms of baseline behavioral assessments, a somatosensory measure (finger proprioception assessed robotically) best predicted treatment gains, outperforming all measures of motor behavior. When the neural basis underlying variability in treatment response was examined, somatosensory-related variables were again the strongest predictors. A multivariate model combining total sensory system injury and sensorimotor cortical connectivity (between ipsilesional primary motor and secondary somatosensory cortices) explained 56% of variance in treatment-induced hand functional gains (<i>p</i> = 0.002). <h3>Conclusions</h3> Measures related to the somatosensory network best explained interparticipant differences in treatment-related hand function gains. These results underscore the importance of baseline somatosensory integrity for improving hand function after stroke and provide insights useful for individualizing rehabilitation therapy. <h3>ClinicalTrials.gov identifier:</h3> NCT02048826.