In vitro cultivated stem cell populations are in general heterogeneous with respect to their expression of differentiation markers. In hematopoietic progenitor populations, this heterogeneity has been shown to regenerate within days from isolated subpopulations defined by high or low marker expression. This kind of plasticity has been suggested to be a fundamental feature of mesenchymal stem cells (MSCs) as well. Here, we study MSC plasticity on the level of individual cells applying a multi-scale computer model that is based on the concept of noise-driven stem cell differentiation.By simulation studies, we provide detailed insight into the kinetics of MSC organisation. Monitoring the fates of individual cells in high and low oxygen culture, we calculated the average transition times of individual cells into stem cell and differentiated states. We predict that at low oxygen the heterogeneity of a MSC population with respect to differentiation regenerates from any selected subpopulation in about two days. At high oxygen, regeneration becomes substantially slowed down. Simulation results on the composition of the functional stem cell pool of MSC populations suggest that most of the cells that constitute this pool originate from more differentiated cells.Individual cell-based models are well-suited to provide quantitative predictions on essential features of the spatio-temporal organisation of MSC in vitro. Our predictions on MSC plasticity and its dependence on the environment motivate a number of in vitro experiments for validation. They may contribute to a better understanding of MSC organisation in vitro, including features of clonal expansion, environmental adaptation and stem cell ageing.