Skip to content

Instantly share code, notes, and snippets.

@djnavarro
Created April 16, 2020 20:42
Show Gist options
  • Save djnavarro/075f71586d9886ecfd99ff6740768a49 to your computer and use it in GitHub Desktop.
Save djnavarro/075f71586d9886ecfd99ff6740768a49 to your computer and use it in GitHub Desktop.
toy example with leaky accumulator models
# illustrate that the temporally-accrued evidence in an Ornstein-
# Uhlenbeck model of perceptual decision-making is bounded
t <- 500 # length of time to run the process
drift <- .2 # drift rate measures the perceptual signal strength
leak <- .9 # evidence leakage rate
# momentary samples are presumed to generate normally distributed
# evidence with mean corresponding to drift rate, and (for reasons
# known but to Roger) standard deviation .1
samples <- rnorm(t, mean = drift, sd = .1)
# discrete-time approximation to the Ornstein-Uhlenbeck model
evidence <- purrr::accumulate(
.x = samples, # <-- my idiot brain REALLY wants this to be .y = samples
.f = ~ (leak * .x) + .y
)
# plot
plot(1:t, evidence, type = "l", xlab = "time")
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment