*4.4. Limitations and Perspectives*

When interpreting the results of this study, the following limitations have to be acknowledged. First, the retrospective character and comparatively small sample size restrict the generalizability of the findings. Upcoming studies may include more patients and follow a prospective study design. Second, the finding of lower GMI for language-positive nTMS points when compared to language-negative nTMS points and related interpretation as a potential hint for higher GM density needs further validation. A potential explanation could also be linked to the increased connectivity profile of language-positive nTMS points. Future studies using imaging with higher resolution may provide evidence for our preliminary interpretations. Third, the technique of DTI has its inherent methodological shortcomings, which could lead to aberrant fiber reconstruction and visualization, particularly for crossing or kissing fibers and in the presence of edema, which is commonly observed

in relation to brain tumors [52,53]. Other sequences and tracking algorithms are developed to become applicable in the clinical setting, but good alternatives to conventional DTI and its tractography-based analyses are still not available for clinical routine use [54,55]. Fourth, the tractography maps generated by nTMS-based DTI FT for depiction of the language network need further validation, ideally by intraoperative DES as the gold-standard method. In this context, previous studies have evaluated the agreement between preoperative nTMS language mapping and intraoperative DES [13,16,24]; however, on the subcortical level, such correlation analyses in representative samples are largely missing to date. First evidence of associations between preoperative nTMS-based tractography and intraoperative DES results or surgery-related aphasia has been obtained [20,37]. Upcoming studies may use more sophisticated approaches for diffusion-weighted MRI and fiber tractography, ideally combining it with functional data such as nTMS maps after further confirmatory studies. Furthermore, repeated investigations of the language network by serial nTMS language mappings and tractography after surgery may be of interest to track potential plastic effects as well as associations with aphasia grades on a longitudinal scale.
