fitgrid

fitgrid is a Python library for running linear models on stacks of epochs.

Here’s a quick demo:

1. Import fitgrid and read an epochs table from an HDF5 file:

   import fitgrid
   epochs = fitgrid.epochs_from_hdf('epochs_table.h5')
   

2. Run a regression model, which creates a FitGrid:

   grid = fitgrid.lm(epochs, RHS='stimulus_magnitude + stimulus_type')
   

3. Now all diagnostic and fit information is available as attributes. For example, the betas:

   betas = grid.params
   

   or adjusted \(R^2\):

   rsquared_adj = grid.rsquared_adj