LICENSE
=======

These files are copyright (c) 20010-2014 by Svenja Marx, Gina Gruenhage and Ueli Rutishauser,
Philipps-University Marburg & California Institute of Technology 

This software is provided AS IS AND WITHOUT ANY WARRANTY WHATSOEVER. 
We welcome comments and questions but can not guarantee any support.
Re-Distribution is NOT permited unless permission is obtained from the authors.

If you use this code or publish based on it, please cite our paper:
S. Marx, G. Gruenhage, D. Walper, U. Rutishauser, W. Einhauser. Competition with and 
without priority control: linking rivalry to attention through winner-take-all networks 
with memory. Ann NY Acad Sci (in press, 2015)

This code is based on 
U. Rutishauser, R.J. Douglas. State dependent computation using coupled recurrent networks. 
Neural computation, 21(2):478-509, 2009

The authors:
svenja.marx@physik.uni-marburg.de
gina.gruenhage@campus.tu-berlin.de
urut@caltech.edu


Introduction
============

This set of mat-files is created to simulate rivalry behavior. The results in 
Marx et al, 2015 (ANYAS) are based on these functions. The basic functions are 
provided here, for testing Levelt's propositions, they have to be run several times
using different input strengths.

 
Functions
=========

MainDoubleWTARivalry.m
Main function, arguments are noise amplitude for the two input units, 
noise offset for the two input units, number of integration steps, noisetype 
(1 = gaussian, else poisson), and blanking table (if necessary).

runDoubleWTA.m, plotDoubleWTAMatrices.m and plotDoubleWTARun_rivalry.m are executed 
in the main file.

test_WTA_blank.m
Function for testing blanking behavior, blanking and presentation duration 
can be adjusted as well as input strengths