Parameter Recovery Studies Individual Model

Parameter recovery I and II. In both studies, we conducted two sets of simulations where we generated stop-signal data from the horse-race model using the ex-Gaussian as-sumption and varied the number of go and stop-signal trials, with 130 synthetic data sets for each set. The estimated µstop, σstop, and τstop parameters were constrained to be equal across five SSDs. Figure 1- Figure 4 show the recovery of the go and stop parameters and the corresponding go RT and SSRT distributions as a function of the number of go and stop-signal trials. The go parameters are recovered very well. The stop parameters are recovered adequately as well, but the posterior standard deviations and the standard errors are quite sizable for the simulations with very few stop-signal trials. Of course, as the number of stop-signal trials increases, the model parameters are estimated more precisely and the standard error decreases. We ran parameter recoveries using various parameter settings; the results are very similar to the ones reported here. Parameter recovery III. Figure 6 shows changes in the estimated go and stop param-eters as a function of changes in the underlying true parameters. We simultaneously varied the value of the generating µgo and µstop parameters from very low to very high in 50 equa