Unravelling the high-dimensional structure of spatial neglect and visuospatial attention: A multivariate approach to lesion-behaviour mapping

Abstract

One of the most studied and elaborated neurological disorders after stroke is probably spatial neglect, a disorder of spatial exploration, attention and awareness occurring in about two third of all right hemispheric stroke patients. A characteristic symptom these patients show is a failure to orient or respond to information on the contralesional side of space including a general orientation to the ipsilesional side. It is still not possible to come to a common consensus regarding this syndrome on theoretical, anatomical and behavioural aspects. The investigation of the anatomical substrates of spatial neglect, however, offers chances to shed light on crucial pathophysiological processes and inform theoretical models. Therefore, a complete research field dedicated several decades of research to the question where in the brain the syndrome of spatial neglect might have its´ pathogenesis and how this information can help us to understand cognitive processes of normal spatial exploration and attentional processing. A method which largely contributed to this field is called lesion-behaviour mapping by drawing statistical inference about the functional brain architecture from focal brain damage. Following the development within the last five to ten years, a new era of computerised lesion-behaviour mapping techniques became widely available, allowing to reiterate and challenge previous findings and to account for the high-dimensional information present in brain lesions. In my thesis I employed these new techniques to unravel the anatomical substrates of the syndrome of spatial neglect and related spatial attentional deficits. I want to show that these methods can be deployed to make valuable contributions to the understanding of the pathophysiology of the syndrome. In my first empirical work, the presence of a large right-hemispheric network related to the behavioural severity of spatial neglect can be confirmed, closing longstanding controversies. It shows that multivariate machine-learning based lesion-behaviour mapping techniques are particularly suited to detect critical brain areas and to evaluate the predictive performance of underlying statistical models. In the second and third empirical work, I complemented these primary findings by applying the same statistical methodology to parameters of remote disconnection and to different diagnostic tools in the assessment of spatial neglect. These works show crucial areas and anatomical hubs severely disconnected to other areas of the brain and contributing to the development of lateralised deficits in spatial neglect patients. Finally, with the last empirical work, contributions to controversial views concerning the anatomical substrates of the extinction phenomenon, a further spatial attentional deficit, were made. By evaluating lesion-behaviour relationships in spatial neglect, as it was done in the present thesis, it will become possible to inform clinical staff how to direct patients to more effective management and treatment schedules, essential for rehabilitation, while spatial neglect generally is considered as a negative prognosis factor for stroke recovery.