Paper link: https://www.sciencedirect.com/science/article/pii/S1053811924000545
(a) Generative Process:
(b) Regularized Discriminative Process:
After the generative process (a) learns the joint latent neurocognitive variables (Section \ref{gen}), the regularized discriminative process (b) retrofits its hierarchical latent space to the joint latent space (Section \ref{disc}). Inference networks
To install the dependencies, you can run in your terminal:
pip install -r requirements.txt
Experimental dataset can be downloaded at: https://zenodo.org/records/8381751
The code is structured as follows:
preprocessing.ipynb
preprocessed the downloaded dataset.data.py
contains functions to transform and feed the data to the model.models.py
defines deep neural network architectures.utilities.py
has utilities for evaluation and plottings.train.py
is the main entry to run the training process.evaluation.ipynb
runs the evaluation.
If you find this code helpful, please cite our paper:
@article{vo2024deep,
title={Deep latent variable joint cognitive modeling of neural signals and human behavior},
author={Vo, Khuong and Sun, Qinhua Jenny and Nunez, Michael D and Vandekerckhove, Joachim and Srinivasan, Ramesh},
journal={NeuroImage},
pages={120559},
year={2024},
publisher={Elsevier}
}