129 Etiology in Lesion-Symptom Mapping: Tumor vs. Stroke in the tumor group. Lesion volume was skewed to larger volumes in the tumor group, and to smaller lesions for stroke. These differences in lesion distribution, overlap and volume had repercussions for the statistical power throughout the brain and consequently the ability to compare LSM; only one-fifth of voxels survived the minimal power threshold for direct comparison. Lesion-symptom mapping In both samples cognitive impairments were present in more than half of the patients. Despite the clear differences in lesion distribution between tumor and stroke, the cognitive profile was quite similar on a group-level. Importantly, there were no significant group-level differences in performance on the memory and fluency tasks. Lesion volume was correlated with cognitive performance in the tumor group. However, the larger lesion volume in the tumor group did not result in more cognitive impairments compared to the stroke population. Memory and fluency tasks and their neuroanatomical correlates present differentiated cognitive processes, as expected. That is, more resemblance was found in critical neuroanatomical locations within than between cognitive domains. For both populations the overlap between the tasks was mainly located in white matter pathways which have been independently related to memory and fluency performance previously.24,46–50 As these white matter tracts have also been related to other cognitive processes, like attention, processing speed and executive functioning,24,42,51 it is most likely that these overlapping regions are involved in a wide range of cognitive processes, including memory and fluency. Of note, overlapping lesions for the tumor and stroke group were mostly found in the white matter and subcortical areas. Thereby the power to find overlapping LSM results between our groups was highest in these white matter areas. In addition, lesion volume correction was performed to eliminate the effect of differences in lesion size between the groups on LSM results as much as possible. In the uncorrected LSM we mainly found larger areas associated with task performance for the tumor group, with significant voxels expanding from the regions found with the corrected LSM. Nevertheless, the overlap between the tumor and stroke LSM results only changed slightly for the uncorrected LSM results, indicating lesion volume control was not a great influence in our comparison between etiologies. Regions associated with memory performance in the tumor group were in both the grey and white matter areas surrounding the left hippocampus. These critical memory-regions are known from literature as part of the core network for episodic memory postulated by Benoit and Schachter52 as well as the human memory circuit from Ferguson and colleagues.53 The left hippocampal areas had insufficient 5
RkJQdWJsaXNoZXIy MTk4NDMw