The RLVM is a latent variable model developed to study large populations of simultaneously recorded neurons.
For more information regarding the mathematical formulation of the model, see the preprint on bioRxiv. The doc
directory contains scripts that show how to use the model on several simulated datasets.
The RLVM optimizes model parameters using Mark Schmidt's minFunc package, which is located in the lib
directory and should work out of the box. If not you may need to run the mexAll.m file from the lib/minFunc_2012
directory.
This is the branch for version 2.0, which contains the following new features:
- autoencoder can now contain an arbitrary number of layers
- refactoring of model structure makes managing large models easier
- stimulus subunits can now be shared across cells
v2.0 is still not fully tested or documented. See the master branch for the most recent stable version.