Categorical encoding of decision variables in orbitofrontal cortex

Categorical encoding of decision variables in orbitofrontal cortex

These are the data and code accompanying the PLoS Computational Biology article "Categorical encoding of decision variables in orbitofrontal cortex" by Arno Onken, Jue Xie, Stefano Panzeri and Camillo Padoa-Schioppa.

Code

The code in the 'code' folder is written in Python and tested with version 3.6.6. The following additional packages are required:
- NumPy tested with version 1.16.5
- SciPy tested with version 1.1.0
- Scikit-learn tested with version 1.19.1
- spherecluster https://pypi.org/project/spherecluster/0.1.2/
- diptest https://github.com/alimuldal/diptest

To reproduce the figures, run the shell script 'plot_figures.sh'. The 'plot_*.py' scripts use the result data in the folders that are specified in the shell script. To reproduce the result data, use the appropriate 'calculate_*.py' script.

Data

The data are located in the 'data' folder.
- 'ofc_psth' contains the peri-stimulus time histograms for each cell. These are used for the PAIRS analysis only (corresponding script 'calculate_pairs.py'.
- 'ofc_rates_all' contains the rates for each cell and trial type in the files 'rates.mat' and a list of the trial types in 'trialtypes.mat'. The other files in the folder are result files generated by the corresponding 'calculate_*.py' scripts.
- 'ofc_rates_shuffled' like 'ofc_rates_all' but the rates where shuffled (see supplementary figure S2 Fig).
- 'ofc_rates_time_windows' contains sub folders for each time window. Each of these sub folders has the same structure as 'ofc_rates_all'.
- 'synthetic_banana' contains the synthetic data used for Figure 3.
- 'synthetic_uniform' contains the synthetic non-categorical data sampled uniformly from the hypersphere (see Figures 5 and 6).
- 'synthetic_variable' contains the synthetic categorical data (see Figures 4, 5 and 6).