DhW: Decoding the hidden Waves

Decoding the hidden Waves: Revealing the Nature of Brain Activity

Riccardo Tedoldi

Supervisor: Uri Hasson

Artificial and Biological Neural Systems, Spring 2023

This repository contains the code for the seminar paper "Decoding the hidden Waves: Revealing the Nature of Brain Activity" (DhW).

Introduction

In this study, I investigated the representation of M/EEG in a feature space. The mapping learned by the networks. It has been shown to cluster the neural activity of correlated event types. Correlation between neural activity and people's thoughts/event type. It would enable modern BCIs to assist people with disabilities.

Installation

Clone the repository

git clone https://github.com/r1cc4r2o/DhW.git
cd DhW

I report here the file to create a conda environment with all the requirements.

conda create --name <env-name> --file ./requirements.txt

Then, you can activate the environment with:

conda activate <env-name>

Datasets

I release the data used in the experiments in a gDrive folder. The data that you can find there have been preprocessed and are ready to be used. You can find the data in both the version: with_pca and without_pca. The data are in the form of torch.tensor and are ready to be used.

import torch

PATH_1 = <path>+'/X.pt'
PATH_2 = <path>+'/y_label_face.pt'
PATH_3 = <path>+'/x_face_embedding.pt'

# to load the data
x = torch.load(PATH_1)
y = torch.load(PATH_2).to(torch.float32)
emb_y = torch.load(PATH_3).to(torch.float32)

y.shape, x.shape, emb_y.shape

Usage

Here the structure of the repository:

├── experiments
│   ├── 0_<name_notebook>
│   ├── 1_<name_notebook>
│   ├── ...
│   ├── n_<name_notebook>
├── src
│   ├── architecture1
│   |   ├── model.py
|   |   ├── _all_the_modules.py
|   |   ├── _utils.py
│   |   ├── train.py # exemplificative
│   ├── architecture2
│   |   ├── model.py
|   |   ├── _all_the_modules.py
|   |   ├── _utils.py
│   |   ├── train.py # exemplificative

The notebook in the folder experiments are numbered and have a name. Then in the folder you can find a README.md file with a short description for each notebook.

Contributing

If you want to contribute to this project, I am happy to hear from you.

Please contact me:

• @riccardotedoldi

Acknowledgements

I thanks Uri Hasson for the supervision and the support during the project. Also, I thanks Christian Brodbeck to help me to fix an issue with the data preprocessing.

To cite this work

@misc{tedoldi2023decoding,
      title={Decoding the hidden Waves: Revealing the Nature of Brain Activity}, 
      author={Riccardo Tedoldi},
      year={2023},
      url  = {https://github.com/r1cc4r2o/DhW}
}