OBJECTIVES: This study was designed to clarify whether facial nerve motor evoked potentials (FNMEPs) elicited by transcranial electrical stimulation during skull base surgery are useful for predicting long-term facial nerve function. METHODS: We analyzed FNMEP findings in 35 patients with skull base tumors. Mean follow-up was 24.4 months. Corkscrew electrodes positioned at C3 or C4 and Cz were used to deliver supramaximal stimuli. FNMEPs were recorded from the orbicularis oculi and oris muscles. RESULTS: The correlation between the final-to-baseline FNMEP ratio and initial or long-term facial nerve function was examined. Initial post-operative facial nerve function correlated significantly with the FNMEP ratios in the orbicularis oculi (r = -0.53, P < 0.005) and orbicularis oris (r = -0.80, P < 0.001) muscles. The correlations between FNMEP ratios and facial nerve function remained significant during long-term follow-up (orbicularis oculi muscle: r = -0.43, P < 0.05; orbicularis oris muscle: r = -0.71, P < 0.001). All patients in whom the FNMEP ratio in the orbicularis oculi muscles remained above 50% were assigned to the satisfactory facial nerve function (House-Brackmann Grades i and ii) group at the final examination. DISCUSSION: FNMEP monitoring can predict facial nerve function not only immediately after surgery but also long-term.