Magnetoencephalography (MEG) is a relatively modern neuroimaging technique able to study normal and pathological brain functioning with temporal resolution in the order of milliseconds and adequate spatial resolution. Although its clinical applications are still relatively limited, great advances have been made in recent years in the field of dementia and Alzheimer's disease (AD) in particular. In this chapter, we briefly describe the physiological phenomena underlying MEG brain signals and the different metrics that can be computed from these data in order to study the alterations disrupting brain activity not only in demented patients, but also in the preclinical and prodromal stages of the disease. Changes in non-linear brain dynamics, power spectral properties, functional connectivity and network topological changes observed in AD are narratively summarized in the context of the pathophysiology of the disease. Furthermore, the potential of MEG as a potential biomarker to identify AD pathology before dementia onset is discussed in the light of current knowledge and the relationship between potential MEG biomarkers and current established hallmarks of the disease is also reviewed. To this aim, findings from different approaches such as resting state or during the performance of different cognitive paradigms are discussed.Lastly, there is an increasing interest in current scientific literature in promoting interventions aimed at modifying certain lifestyles, such as nutrition or physical activity among others, thought to reduce or delay AD risk. We discuss the utility of MEG as a potential marker of the success of such interventions from the available literature.