Simulate from a go/no-go diffusion model

Draws from the generative process: run the two-boundary diffusion, and record a response only if it reached the upper boundary before the deadline. Nothing here evaluates the go density or the no-go probability, so the simulator is an independent statement of the same model and can be used to check the likelihood rather than merely to agree with it.

Usage

wiener_gng_simulate(
  n,
  mu = 1,
  bs = 1.4,
  ndt = 0.25,
  bias = 0.5,
  deadline = 1.5,
  sv = 0,
  sz = 0,
  st = 0
)

Arguments

n

Number of trials.

mu

Drift rate.

bs

Boundary separation.

ndt

Non-decision time.

bias

Relative start point in (0, 1).

deadline

The response deadline. One number, or one per trial.

sv, sz, st

Across-trial variability in the drift rate, the relative start point and the non-decision time, as ddm_simulate() takes them. Zero, the default, is the plain model.

Value

A data frame with responded (1 for a go trial), rt (the response time on a go trial, and the deadline on a no-go trial, which is the placeholder wiener_gng() documents) and deadline.

Examples

set.seed(1)
dat <- wiener_gng_simulate(1000, mu = 1.0, bs = 1.4, ndt = 0.25,
                           deadline = 1.5)
mean(dat$responded)
#> [1] 0.759
summary(dat$rt[dat$responded == 1])
#>    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
#>  0.2950  0.4431  0.5709  0.6473  0.7767  1.4981