README.md

In rettopnivek/nROUSE: Functions for the nROUSE model

nROUSE

The R package nROUSE provides a set of functions to simulate and estimate the nROUSE model (Huber & O'Reilly, 2003; Rieth & Huber, 2017). The nROUSE model is a neural network model that accounts for accuracy performance in experimental paradigms that involve perceptual identification with repetition priming. The model posits that manipulations of prime type and duration influence accuracy performance via the non-linear interaction of lingering perceptual activation and neural habituation.

Getting started

Prerequisites

• The program R ( version >= 3.0.0 )
• A C++ compiler (e.g., via Rtools)
• The R packages Rcpp and RcppArmadillo (for the underlying C++ code and matrix algebra).

Installation

To easily install the 'nROUSE' package, you'll need the 'devtools' package:

install.packages("devtools")
library(devtools)

The 'nROUSE' package can then be installed via the following command:

install_github("rettopnivek/nROUSE")

Using the package

The package can be loaded via:

library(nROUSE)

A figure with a visual demonstration of how the nROUSE model generates predictions via simulated neural activity can be produced with the command:

nROUSE_demonstration()

The predicted accuracy over varying prime durations and types can be produced via:

nROUSE_predictions()

The nROUSE model can be simulated via simulate_nROUSE, and the log-likelihood for a set of data can be computed via the function nROUSE_logLik. This latter function can be passed into a optimization routine (e.g., optim) to carry out maximum likelihood estimation.

Additional details and examples for all four functions can be obtained via:

help("function")

substituting the appropriate function name for "function".

Authors

Kevin Potter

License

MIT

rettopnivek/nROUSE documentation built on May 27, 2019, 5:55 a.m.