Hippocampal theta oscillations are correlated with complex behaviours such as memory and navigation. Measuring this activity non-invasively in humans is attractive but challenging – with magnetoencephalography (MEG) both the hippocampus’ depth and the spatial uncertainty introduced during data acquisition and co-registration constitute obstacles. In this talk, I will describe how the depth can be accounted for through anatomical modelling, and the spatial uncertainty can be minimized by using head-casts. I will then discuss results from an experiment combining the anatomical modelling, head-casts, and a spatial memory virtual reality paradigm. I hope to demonstrate that this set-up enables temporally resolved, non-invasive, in vivo recordings of human hippocampal activity using MEG. Within this framework we are able to begin to replicate and potentially extend findings from invasive human and rodent literature. Finally, I will show new data recorded with non-cryogenic MEG sensors which can be placed directly on the scalp, describe how an array might be optimised to record hippocampal signals, and discuss the potential for these sensors to replace conventional cryogenic MEG.