Objective: MEG and EEG after sleep deprivation (EEG-SD) are applied as diagnostic tools in the evaluation of patients with possible epilepsy. There is no gold standard to check whether the diagnosis based on these two modalities is correct. The best standard available is the long-term follow-up of patients. As follow-up of an earlier study in which the additional value of MEG vs EEG-SD diagnosis was evaluated, we investigated the long-term validity of MEG-based and EEG-SD-based diagnosis. Materials and methods: Data collected from 46 patients were used in a comparative study of the last known diagnosis against the original one of 8 years ago. Results: Long-term (3–8 years) sensitivity of sharp phenomena (combining spikes and sharp waves) in routine MEG and in EEG-SD for the diagnosis epilepsy is 71% and 62%, respectively. When compared to the original study, this hardly changed. Over time, uncertainty on diagnosis diminishes. Conclusion: MEG as well as EEG-SD are robust long-term predictors for epilepsy.