rlba: Simulate Choices and Response Times

In lbaModel: The Linear Ballistic Accumulation Model

Simulate Choices and Response Times

Description

Simulate Choices and Response Times

Usage

rlba_r(
  parameter_r,
  is_positive_drift_r,
  time_parameter_r,
  n = 1L,
  use_inverse_method = FALSE,
  debug = FALSE
)

rlba(
  parameter_r,
  is_positive_drift_r,
  time_parameter_r,
  n = 1L,
  use_inverse_method = FALSE,
  debug = FALSE,
  seed = NULL
)

Arguments

parameter_r	A numeric matrix of LBA parameters. Each column corresponds to an accumulator, and each row corresponds to a parameter, with the expected order being: A - Starting point variability b - Thresholds mean_v - Mean drift rates sd_v - Standard deviation drift rates st0 - Variability of the non-decision t0 - Non-decision time
is_positive_drift_r	A logical vector indicating whether the drift rate (v) is strictly positive.
time_parameter_r	A numeric vector to set the simulated time grid, with the expected value for: minimal and maximum decision times and the difference time (i.e., min dt, max dt and dt).
n	Integer. Number of trials to simulate. Defaults to 1L.
use_inverse_method	Logical. If TRUE, use the inverse transform sampling method; otherwise use the standard sampling method.Defaults to FALSE. TODO: fix the memory problem.
debug	Logical. If TRUE, prints detailed messages for debugging. Defaults to FALSE.
seed	Optional integer. If provided, sets the random seed for reproducible simulation. Defaults to NULL. Available only in the R interface.

Details

This function generates simulated data for the LBA model. The function supports both standard and inverse transform sampling methods. The inverse method may offer better numerical behaviour in edge cases (e.g., small RTs or boundary conditions).

Value

A data.frame. Each row corresponds to a simulated trial containing:

• trial — Trial index

• choice — Index of the winning accumulator

• rt — Simulated response time

Examples

param_list2mat <- function(param_list) {
  n_row <- length(param_list[[1]])
  n_col <- length(param_list)
  out <- matrix(NA, nrow = n_row, ncol = n_col)

  for (i in seq_len(n_col)) {
    out[, i] <- param_list[[i]]
  }
  t(out)
}

A <- 1.2
b <- 2.7
t0 <- .2

mean_v <- c(2.4, 2.2)
sd_v <- c(1, 1)

RT <- seq(0, 3, .4) + t0
posdrift <- TRUE
nv <- length(mean_v)

st0 <- 0

params_tmp <- list(
  A = rep(A, nv),
  b = rep(b, nv),
  mean_v = mean_v,
  sd_v = sd_v,
  st0 = rep(st0, nv),
  t0 = rep(t0, nv)
)

params <- param_list2mat(params_tmp)
is_positive_drift <- rep(TRUE, nv)

n <- 1
seed <- 123

dt <- 0.01
min_dt <- 0
max_dt <- 5
time_parameter_r <- c(min_dt, max_dt, dt)


set.seed(seed)
# R interface
result1 <- rlba(params, is_positive_drift, time_parameter_r, n,
  seed = seed, debug = TRUE
)
# C++ interface. No seed argument.
result2 <- rlba_r(params, is_positive_drift, time_parameter_r, n,
  debug = TRUE
)

print(result1)
print(result2)

lbaModel documentation built on Sept. 15, 2025, 9:08 a.m.