147 SUMMARY & GENERAL DISCUSSION 7 use our current findings to look at how we could improve rehabilitation. As mentioned in chapter 6, the specific rehabilitation program we investigated was not developed to specifically target visuomotor processes we now know to be involved in NA. The program aims to improve scapular stability and coordination through coordinative training targeting central motor control, which may include use of visual feedback (through mirror or video) but does not actively train visuomotor processes related to processing of selfrelevant information or sensorimotor integration. We could potentially improve specific rehabilitation for NA patients by applying visuomotor rehabilitation strategies that have shown promising results in other nervous system disorders. In this thesis we used motor imagery to assess sensorimotor representations in NA. Motor imagery practice involves explicit imagination of movements and has been shown to positively affect clinical outcome in musculoskeletal disorders, 181 and both clinical and cerebral outcome in upper extremity stroke rehabilitation and arm immobilization. 224-226 Action observation, often combined with motor imagery practice, has shown similar clinical effects in Parkinson’s disease, multiple sclerosis, cerebral palsy, and orthopaedic diagnoses, 180, 227, 228 as well as cerebral and clinical effects in stroke. 229, 230 Graded motor imagery, which combines mental hand rotation, explicit motor imagery and mirror visual therapy, can reduce phantom limb pain in amputees. 231 It has also yielded promising clinical results in complex regional pain syndrome patients showing improvements on motor and sensory domains, along with changes in related brain activity. 232, 233 Graded motor imagery may therefore be especially useful for NA patients that experience pain. There are indications that mirror visual feedback on its own, can improve clinical outcome and influence brain activity in a variety of (neurological) disorders, by promoting integration of different sensory modalities. 179, 230, 234 Although visual feedback is already part of the specific rehabilitation program, future work could explore how such strategies could be further incorporated. Recent neuroimaging studies in healthy volunteers suggest that motor imagery practice effects can be further enhanced by applying neurofeedback235 and immersive virtual reality, 236 although the clinical application of the former may be hindered by the required resources. Virtual reality has also been shown to improve clinical outcome in upper extremity stroke rehabilitation. 237 These and other strategies have the potential to target visuomotor processes and improve sensorimotor integration in NA patients, and ultimately improve clinical outcome. Taken together, we can conclude that although much work is still needed to identify what, when, which, and how to improve rehabilitation after NA, our findings and existing literature provide us with important leads to move forward. Concluding remarks With this thesis, we provide empirical evidence confirming the clinical intuition that the brain adapts to peripheral nerve injury in NA and that these adaptations are associated with clinical outcome. Maladaptation occurs in visuomotor brain areas and can recover towards normality alongside clinical recovery. This offers crucial new insights in NA disease mechanisms and importantly opens up potential new avenues for treatment of residual complaints in this patient population. Specific, multidisciplinary, outpatient rehabilitation can aid clinical recovery, but can be further optimized to target cerebral processes and improve clinical outcome. In order to reach this goal, we need to gain a better understanding of the role cerebral mechanisms play in recovery and rehabilitation in NA. In order to do so, we need to determine: