131 Etiology in Lesion-Symptom Mapping: Tumor vs. Stroke Though power inequalities could explain some of the differences, post-hoc analyses in regions with adequate coverage in both groups confirmed that the results of lesion-deficit inference strongly differ. To illustrate, in several brain areas the negative impact of a lesion in this area on semantic fluency performance was only present when this brain area was lesioned by a tumor. This was, however, not the case for all investigated brain areas. For example, all post-hoc analyses in brain regions that were significantly associated with cognitive performance in the stroke group, did not show a similar etiology interaction effect. Here, differences might be better explained by the large variation in cognitive performance, thereby failing to reach significance in the lesion-symptom mapping analyses. We specifically investigated the effect of etiology on lesion-symptom associations identified by lesion-symptom mapping. In both tumor and stroke, however, the impact of a lesion can extend far beyond local changes in circumscribed tissue visible on standard imaging techniques, and a small lesion can have widespread effects on behavior.8,28,57–61 This might also explain the similarity in cognitive profiles between the two samples despite clear differences in lesion location and volume. It is likely that the degree of overlap between both samples is underestimated in terms of structural and functional connections. Nevertheless, this does not explain why we found different cognitive effects in brain areas with overlapping lesion coverage. Etiology-specific cognitive effects Few studies discuss the impact of etiology on lesion-behavior associations. Anderson and colleagues matched a small group of brain tumor patients (n = 17) to patients with unilateral stroke based on lesion location and compared cognitive outcome.62 They showed substantial differences in cognitive performance, with tumor patients outperforming matched stroke subjects and showing relatively mild impairments. Accordingly, it is still often assumed that the location of a tumor bears little explanatory value. This assumption was challenged by Shallice and colleagues63, who did find selective and dissociated visuo-spatial deficits in a series of brain tumor patients. Two studies of more recent date directly compared the impact of different etiologies on the cognitive sequelae of lesions. Frontal-based functions were compared between tumor and stroke patients and showed no differences in performance.16 Here, the authors selected patients with frontal lesions, but did not perform further LSM. In a LSM study of apraxia, tumor and stroke patients were included and analyzed both separately and combined.64 Critical brain areas for apraxia matched previous literature, but differed in the tumor and stroke group. This is in line with 5
RkJQdWJsaXNoZXIy MTk4NDMw