Human empathic experience is a multifaceted psychological construct which arises from functional integration of multiple neural networks. Despite accumulating knowledge about the cortical circuitry of empathy, almost nothing is known about the connectivity that may be concerned in conveying empathy-related neural information. To bridge this gap in knowledge, we studied dispositional empathy in a large-sized cohort of 107 patients who had undergone surgery for a diffuse low-grade glioma. The self-report questionnaire used enabled us to obtain a global measure of subjective empathy but also, importantly, to assess the two main components of empathy (cognitive and emotional). Data were processed by combining voxelwise and tractwise lesion-symptom analyses. Several major findings emerged from our analyses. First of all, topological voxelwise analyses were inconclusive. Conversely, tractwise multiple regression analyses, including all major associative white matter pathways as potential predictors, yielded to significant models explaining substantial part of the behavioural variance. Among the main results, we found that disconnection of the left cingulum bundle was a strong predictor of a low cognitive empathy (p<0.0005 Bonferroni-corrected). Similarly, we found that disconnection of the right uncinate fasciculus and the right inferior fronto-occipital fasciculus predicted, respectively, a low (p<0.05 Bonferroni-corrected) and a high (p<0.05 Bonferroni-corrected) subjective empathy. Finally, although we failed to relate emotional empathy to disruption of a specific tract, correlation analyses indicated a positive association between this component of empathy and the volumes of residual lesion infiltration in the right hemisphere (p<0.01). Taken as a whole, these findings provide key fundamental insights into the anatomical connectivity of empathy. They may help to better understand the pathophysiology of empathy impairments in pathological conditions characterized by abnormalities of long-range anatomical connectivity, such as autism spectrum disorders, schizophrenia and fronto-temporal dementia 1).