The Context-Unified Encoding (CUE) model

A spiking neural network model extending the ordinal serial encoding (OSE) model and temporal context model (TCM) to integrate short- and long-term memory.

Installation

git clone https://github.com/ctn-archive/cue-model.git
cd cue-model
pip install .

Requirements

• Python 3.5 or later (earlier versions might work, but are untested)

Further dependencies should be installed automatically during the installation:

• Nengo 2.5 or later
• nengo_spa 0.3.x
• nengo_extras
• matplotlib
• NumPy
• SciPy
• Pandas
• PyTry
• Psyrun
• SciPy
• Seaborn
• statsmodels

Running the model

Single trials can be run with the pytry command and one of the files in cue/trials as argument. For example:

pytry cue/trials/default.py --data_format npz --seed 42

Use the --help argument to get a list of command line options and parameters:

pytry cue/trials/default.py --help

To run larger sets of simulations with different seeds use the psy run command. Without arguments it will run multiple simulations for all experimental conditions. To run only selected experimental conditions, pass them as arguments. For example:

psy run immediate delayed

To list available experimental conditions use psy list.

Data evaluation and plots

For data evaluation and plots see the Jupyter notebooks in the notebooks directory.