Fig 4. Within a few hundred iteration steps, our parameter estimates converge to the correct theoretical value whether we use the exact (Eq (9)) or approximate (Eq (10)) likelihood function.

Iteration steps are the ΔT’s along the trajectories used to make the point estimate of our parameters. In A) and B) we consider a point source without memory (m = 0) and with memory (m = 5), respectively. The results of these calculations confirm that our first cumulant approximation of the likelihood function (shown in Eq (10)) eventually converges to the correct theoretical parameter values. In C) we show the same results in the absence of a gradient using the exact number of hits.