Author Summary
Molecules and receptors in the cell membrane undergo complicated diffusion motion due to the intricate
architecture of the cell membrane. We recently demonstrated that both interactions and diffusion maps
could be inferred from single molecule trajectories. A major goal of tracking species in the cell membrane
is to better understand how they interact with cell membrane features, such as corrals, domains and
fences. Such data carries a lot of information on the biological function of the tracked species. We
develop a method based on Bayesian statistics that addresses the question of reliably quantifying how a
tracked single molecule moves in the cell membrane. We define and test the performance of three criteria,
which we have adopted from the field of information theory. With these criteria, we build a decision
tree that helps to determine if a tracked molecule or receptor undergoes free Brownian motion or if it is
confined. The technique can correctly classify experimental data obtained from a toxin receptor. This
method adds another dimension to data analysis of such trajectories.