We present a synchronized multimodal neuroimaging dataset for studying brain language processing (SMN4Lang) that contains fMRI and MEG data on the same 12 healthy volunteers while the volunteers listened to 6 hours of naturalistic stories.