Research Group Leader

Prof. Dr. Matthias Bethge
Prof. Dr. Matthias Bethge
Phone: +49 7071 29-89017
Fax: +49 7071 29-25015


Secretary: Heike König
Phone: +49 7071 29-89018
Fax: +49 7071 29-25015


Neurometric Function Analysis for Neural Population Codes

This package contains MATLAB functions to replicate the results presented in the papers

  • Berens et al. (in press), Optimal population coding, revisited, Proceedings of the National Academy of Sciences USA
  • Berens et al. (2009), Neurometric function analysis of neural population codes, Advances in Neural Information Processing Systems (NIPS)
The package contains functions for numerically computing the MDE, the IMDE and the MMSE for a homogenous population with responding to an angular variable with a Gaussian Poisson-like noise model. Fisher information and the MASE are also computed. It also contains a function to estimate the MDE of an independent population with discrete Poisson distributed spike counts.
The package contains a demo.m file which explains how to use the functions. In addition, all functions contain comments, which might be helpful for understanding the code.

It should be easy to adapt the code for specific other uses, since it is fairly modular. The core routines are estimateMMSE.m and estimateMDE.m, where the error measures for a specific population are computed. If you require other tuning function shapes or covariance structures, you need to change genPop.m, genMean.m and genCov.m.

If you discover problems or - worse, but inevitable - a bug, please send me an email.

License: Creative Commons by-nc-sa 3.0 (License)

Last updated: Wednesday, 27.02.2013