Pharmacoresistant mesial temporal lobe epilepsy (mTLE) represents the major indication for epilepsy surgery. Since epilepsy surgery is an elective treatment option, preserving cognition is a high priority. Given the essential role of temporomesial structures in declarative long-term memory formation, surgical treatment for mTLE is primarily associated with a risk of material-specific memory decline, but other cognitive domains may be affected as well. The major determinants for the neuropsychological outcome are the functional integrity of surgically affected tissues, the functional reserve capacities of the remnant brain, the postoperative seizure outcome, as well as the quantitative and qualitative changes of antiepileptic drugs. Anterior temporal lobectomy has long been the standard procedure for treating mTLE. However, if an exclusive mesial pathology is present, then functional non-pathological tissues of the temporal pole and neocortex are sacrificed. As a result, more selective or tailored surgical approaches have been developed which strive towards minimizing iatrogenic effects. However, whether or not these approaches are equipotential with regard to seizure control is still a matter of debate. The quality of the presurgical diagnostics could also be decisive. Selective surgery should indeed be selective in terms of preventing evitable collateral cortical damage along the approach. Invasivity and risks of collateral damage associated with "open" selective resective surgery are further minimized by stereotactic ablative surgery via thermocoagulation, or eventually even eliminated by gamma knife surgery. From a neuropsychological point of view, this development is consequent and desirable, but no clear scientific evidence of a superior cognitive outcome after radiosurgery or thermocoagulation currently exists. The studies that are available contain significant methodological limitations. Thus, randomized head-to-head cognitive outcome studies of competing selective procedures are needed, which should meet the minimum requirements for study design and neuropsychological evaluations. Finally, none of the surgical treatment variants can systematically prevent memory decline when the hippocampus is targeted.