Code for project in which drift depends on dynamical systems.
To run the code in this repo locally, install the following python packages into your enviroment with pip.
Here some code to generate a conda enviroment and install the necessary packages.
conda create --name dynddm python=3.7
conda activate dynddm
conda install pip
conda install jupyter
pip install matplotlib
pip install git+https://github.com/AlexanderFengler/ssm_simulators
You can run the jupyter notebooks in colab, just click on the links below:
The test_simulators.ipynb notebook provides a starting point to play around with the ds_conflict_drift model.
The model assume the following form for the drift:
v(t) = w_target(t) * target_coherence + w_distractor(t) * distractor_coherence
where (equivalently for w_distrator),
w_target(t; initial_position, fixed_point, slope) = (initial_position - fixed_point) * exp(-(slope*t)) + fixed_point
The rest of the model follows the skeleton of a standard drift diffusion model.
Also available is the ds_conflict_angle model. This includes a collapsing bound but is otherwise identical to the ds_conflict_drift model.