Fig 2. Ensemble-averaged statistics are similar across models.

All of the proposed superdiffusive models are capable of capturing ensemble-averaged mouse fibroblast statistics. (A) Mean-squared displacements for mouse fibroblast cells, shown in red, as well as a Lévy walk and a generalized SPP model. Both models are able to match the mouse fibroblast MSD within the margin of error. (B) The velocity auto-correlation function C_vv as a function of time dt. There is a sharp decrease in the VACF across the first frame, due to error in reconstructing the nuclei centers-of-mass generated by imaging noise and fluctuations in dye intensity. At the largest timescales, each bin corresponds to fewer events and so error bars become large. In addition, adding positional error to simulation trajectories to match the initial dropoff in the VACF causes significant fluctuations at larger timescales.