Goodness of Fit Framework for Neural Population Models using Time-rescaling Theorem

Long Tao, Karoline E. Weber, Kensuke Arai, Uri T. Eden

A common goodness-of-fit framework for neural population models using marked point process time-rescaling (2018)

Source code to create Fig. 1, 2, 4 is here

popTRT tool

popTRT rescales spike times for marked spikes modeled with a marked point process model. By default, we provide the kernel based model which can be calculated from the spikes alone, or the user may provide parameters for a mixture of Gaussians model for the conditional intensity function.

Required packages

numpy matplotlib cython

Installing popTRT

python setup.py install

Using popTRT

dataformat

Marked spike data

posx(t) spk01(t) mark1...K

2 + K columns for 1-dimensional position, 0 or 1 spks in time bin and K dim mark (if spike == 1)

Directories

popTRT python and cython files run the python run scripts where one would enter parameters of the run, ie model type, input data file location etc.