136 Chapter 5 whether the cognitive effects of damage to critical white matter structures also differ between etiologies. CONCLUSIONS In this large-scale, direct comparison of LSM of patients with brain tumors versus ischemic stroke, we found substantial differences in lesion volume and topography between the groups. These differences partly drive the brain-behavior relationships that we found. In contrast to our expectations, the overlap of LSM in brain areas affected in both populations was minimal, though we did find white matter tracts involved in memory and semantic fluency performance across etiologies. Posthoc analyses confirmed an interaction effect between lesion status and etiology for most brain areas with adequate coverage in both groups. Thus, we conclude that lesion-behavior associations as defined by LSM, are influenced by the etiology causing the lesion. Our findings cannot solely be explained by previous objections to the use of tumor patients in LSM. The pattern shown by tumor patients on the group level is consistent with localizations found in earlier studies using different techniques. In agreement with Shallice and Skrap80, we argue that tumor series can be used to provide converging evidence about functional localization, next to evidence from other techniques such as functional imaging and direct electrocortical stimulation. In clinical terms, this study suggests that data from ischemic stroke patients have only limited value for the prediction of behavioral repercussions of specific lesions caused by primary brain tumors, and vice versa. Because of the lack of generalizability of findings across etiologies, we are cautious about grouping different etiologies in LSM, because results are easily driven by one population. Instead, we advocate to test predictions based on one etiology in a second patient population and explore divergent findings. It would be interesting, for example, to test to what extent divergent findings between tumor and stroke result from plasticity or preservation of function using cortical mapping observations during awake surgery. SUPPLEMENTARY MATERIALS
RkJQdWJsaXNoZXIy MTk4NDMw