Here, we have optimised pre-surgical language mapping using navigated TMS with advanced E-field modelling and precise stimulation pulse timing. We employed a multi-modal approach incorporating ultra-high field (3 and 7 Tesla) functional magnetic resonance imaging, functional causal mapping including neuronavigated repetitive TMS, a software that allows for well-defined targeting and visual stimulus delay timings, as well as an advanced E-field/behavioural analysis. With this approach we could show that ‘virtual lesions’ close to tumour tissue yields clear spatial maps of functional impairment in language production.