*4.2. White Matter Tractography*

The ratios for fiber volumes of U-fibers, transcallosal fibers, and fibers coursing to the cerebellum were higher for language-positive areas compared to language-negative counterparts. Thus, language-related spots seem to be characterized by an enhanced connectivity profile, which exists focally for short fibers connecting cortical regions between adjacent gyri (increased RUfibers) as well as more remotely for long connecting fibers (increased Rcross and Rcereb). This may reflect either higher a-priori involvement in the cortico-subcortical language network, compensatory mechanisms in light of language function at risk in brain tumor patients, or a mixed picture of both. While the specificity of nTMS language mapping in comparison to intraoperative DES has shown to be comparatively low with 23.8%, the technique's negative predictive value was 83.9%, implicating that language-negative nTMS points are also mostly negative in intraoperative DES mapping [13]. The significantly lower connectivity profile of language-negative nTMS spots as revealed by the present study may resemble the "truly" absent or less involved character of these spots. Furthermore, a previous study on nTMS-based DTI FT revealed that there is a higher likelihood of subcortical connections with language-positive nTMS spots as compared to language-negative ones, with true-positive connections (connections for positive spots) being visualized up to four-fold more frequently than false-positive connections (connections for negative spots) [21]. The clear difference in this likelihood could reflect high reliability of nTMS-based DTI FT for the purpose of tracking parts of the human language network [21]. Likewise, the enhanced connectivity profile to adjacent gyri as well as the contralateral hemisphere and cerebellum may serve as a surrogate of good reliability.
