PAL_AMPM_Classify_Demo: A modification of the psi-marginal method allows user to optimize the classification of an observer into one of multiple populations/categories. In this example, two populations are defined which differ with respect to location and slope parameters of the PF. The two populations are displayed as surfaces in location x slope space (top right figure, note that in the figure 30 trials have been performed and the posterior probability of one of the populations is near zero, which is why this population does not show anymore). Populationns are also shown as PFs with 68% and 95% high-density regions in the bottom figure (note that here the high density regions merely indicate the prevalence in the populations, not the ensuing estimation accuracy). In this example, the method treats location, slope and category/population as free parameters and guess rate and lapse rate as fixed parameters. The user is prompted to either optimize estimation of all parameters or to optimize only with respect to category membership. The latter option will generally lead to improved categorization performance (fewer trials needed to reach any desired categorization confidence level) at the cost of less precision in the estimates of location and slope values.