Aberrant folded proteins and peptides are hallmarks of amyloidogenic diseases. However, the molecular processes that cause these proteins to adopt non-native structures in vivo and become cytotoxic are still largely unknown, despite intense efforts to establish a general molecular description of their behavior. Clearly, the fate of these proteins is ultimately linked to their immediate biochemical environment in vivo. In this review, we focus on the role of biological membranes, reactive interfaces that not only affect the conformational stability of amyloidogenic proteins, but also their aggregation rates and, probably, their toxicity. We first provide an overview of recent work, starting with findings regarding the amphiphatic amyloid-beta protein (Abeta), which give evidence that membranes can directly promote aggregation, and that the effectiveness in this process can be related to the presence of specific neuronal ganglioside lipids. In addition, we discuss the implications of recent research (medin as an detailed example) regarding putative roles of membranes in the misfolding behavior of soluble, non-amphiphatic proteins, which are attracting increasing interest. The potential role of membranes in exerting the toxic action of misfolded proteins will also be highlighted in a molecular context. In this review, we discuss novel NMR-based approaches for exploring membrane-protein interactions, and findings obtained using them, which we use to develop a molecular concept to describe membrane-mediated protein misfolding as a quasi-two-dimensional process rather than a three-dimensional event in a biochemical environment. The aim of the review is to provide researchers with a general understanding of the involvement of membranes in folding/misfolding processes in vivo, which might be quite universal and important for future research concerning amyloidogenic and misfolding proteins, and possible ways to prevent their toxic actions.