143 SUMMARY & GENERAL DISCUSSION 7 their ability to lift their arm. Likewise, it is probable that the support provided by the scanner bed reduced scapular instability, making it easier to perform the (imagined) movement. Other, contextual factors, not specific to NA, that may have reduced motor imagery quality and subsequent sensitivity to cerebral changes include a scanner-induced decrease in attentional focus, 174 and inability to execute the imagined movements due to the constraints imposed by the scanner environment . 206, 207 Future studies could explore alternative neuroimaging techniques, such as functional near-infrared spectroscopy (fNIRS), magnetoencephalography (MEG) or electroetoencephalography (EEG) to reduce the influence of these factors and improve ecological validity. Finally, it might be more difficult to discern changes in the core sensorimotor system in NA compared to other disorders of the nervous system, regardless of the paradigm that is used. The heterogeneity of nerve involvement and severity of the nerve injury is considerable in the NA population. 7 Damage can occur anywhere in the brachial plexus territory (and even beyond), and motor and sensory symptoms vary depending on which nerves are affected and the extent of their injury. 5-7 This heterogeneity may make it more difficult to find group effects in somatosensory and somatomotor regions, compared to other peripheral nervous system disorders with a more homogenous, localised distribution (e.g. carpal tunnel syndrome and median nerve injury which mainly involve the hand and wrist), and/ or more extensive and severe damage (e.g. upper limb amputation or traumatic brachial plexus injury). Nevertheless, we only selected patients that presented with scapular dyskinesia. The somatotopic maps of the shoulder are however considerably smaller and less detailed than those of the hand. 208 In combination with the complexity of shoulder movement, which has many degrees of freedom, it may therefore be more difficult to isolate specific aspects of shoulder movement and assess the underlying somatotopic representations. In the end, each of the factors described above may have played a role in reducing our sensitivity to adaptations in the core sensorimotor system. Future studies could focus specifically on the core sensorimotor system, where possible controlling for these factors, to help better understand its role in residual complaints after peripheral nerve damage in NA. Regardless of the factors outlined above, our pioneering findings hold valuable lessons. Our paradigm enabled us to confirm the central hypothesis of this thesis: that cerebral (mal)adaptations occur in response to peripheral nerve injury in NA, as it allowed us to study cerebral sensorimotor representations while controlling for disease-related peripheral changes in motor execution and related afferent feedback. It appears that the cerebral somatotopic representations of the upper extremity are relatively robust in response to peripheral nerve injury in NA. Instead, (mal)adaptations occur in visuomotor areas involved in integration of information from different sensory modalities. It was not possible to dissociate sensory modalities with the paradigm we used here. It would be of interest to discern how NA patients with persistent motor dysfunction process and integrate input from different sensory modalities, in particular visual and proprioceptive input. This could be achieved through pointing or reaching tasks that manipulate and/ or separate target and feedback modalities147, 209-215 such as adaptation and perturbation paradigms. Such tasks could improve our understanding of the relative role each modality plays in altered sensorimotor processing in NA, which could help further specify targets for treatment. Despite the fact that we are still missing pieces of the puzzle, we can