Full recovery is not expected after peripheral nerve repair in the upper limb. Resulting impairments severely limit the patient's ability to use their injured hand in everyday activities and lead to a large financial burden for the individual and society more widely. Evidence suggests that central factors play a major role in limiting recovery. "Faulty touch localisation" is widely recognised as a hallmark of impairment in human patients and animal models reveal dynamic reorganisation of digit maps in primary somatosensory cortex (S1) after nerve repair. Yet, the applicability of map changes to humans and their functional implications remains unknown. In this thesis we study touch localisation and cortical organisation in 21 patients with repair of the median and/or ulnar nerve.
In chapter 2, we employ a novel method to measure touch localisation in which the participants localise touch by indicating the perceived location of a point stimulus on a photograph of their hand. Consistent with previous literature, we find elevated error of localisation in patients that is limited to the territory of the impaired nerve. Additionally, a few patients show an abnormal amount and pattern of misreferrals - errors made across digits or from the digits to the palm.
In chapter 3, we use functional magnetic resonance imaging (fMRI) to reconstruct the organisation of digit maps in S1 to stimulations of the contralateral hand. Consistent with work in animal models, univariate analysis using Dice overlap coefficients revealed a larger overlap of positive-going blood-oxygenation level dependent (BOLD) activity as well as changes to the structure of digit representations. As expected, the increase in overlap was limited to the territory of the injured nerve. Surprisingly, however, both the cortex contralateral and ipsilateral to the injured hand showed structural changes. Additional multivariate analysis using representational distances showed the same structural changes contralateral and ipsilateral to the injury, but we did not observe any changes to the similarity of the multivariate response patterns. Despite clear cortical reorganization and localisation deficits specific to the injured nerve, direct correlations between the two were not found.
Overall, our results confirm previous qualitative observations with rigorous, quantitative methods and indicate that humans exhibit dynamic cortical plasticity in S1 comparable to animal models. At the same time, they highlight the complexity of the interplay between cortical changes and peripheral impairments. Comprehensive quantification of both aspects within a detailed computational model is needed to understand their intricate relationship and improve functional outcomes.