| --- |
| license: mit |
| --- |
| |
| GENEEG dataset from the paper [On the challenges of detecting MCI using EEG in the wild](https://arxiv.org/abs/2501.17871) |
|
|
| The raw files contain raw EEG values (`.eeg` files) for patients, with corresponding: |
| 1. `.art` files that denote level of artifact at a given time position. 0 -- No noise, 1 and more -- Noise. |
| 2. `.evt` files that denote external stimulus at a given time position. |
|
|
| There is data for 17 channels: FP1, FP2, F3, F4, F7, F8, C3, C4, P3, P4, O1, O2, T3, T4, FZ, CZ, PZ. So the `.eeg` and `.art` files contain 17 space separated columns, while the `.evt` file contains only 1 column denoting the stimulus. |
|
|
| The `.pkl` files contain filtered data from patients as a dictionary of numpy arrays which have been artifact filtered (only 0 noise level), and contig length (200 here) chopped. |
|
|
| The dictionary has two keys "wmci" and "wctrl", representing the two classes MCI and Control for the WAVi dataset. |
|
|
| Each element in the dictionary is a list of patients data stored as a list of arrays. |
|
|
| ``` |
| -wmci |
| |-patient 1 |
| ||-contig 1 (a [17 x 200] array) |
| ||-contig 2 |
| ||... |
| ||-contig N_1 (number of contigs for patient 1) |
| |-patient 2 |
| ||-contig 1 (a [17 x 200] array) |
| ||-contig 2 |
| ||... |
| ||-contig N_2 |
| |... |
| -wctrl |
| |-patient 1 |
| |... |
| ``` |
|
|
| If you find this dataset useful, please cite as: |
| ``` |
| @article{mishra2025challenges, |
| title={On the challenges of detecting MCI using EEG in the wild}, |
| author={Mishra, Aayush and Joffe, David and Telidevara, Sankara Surendra and Oakley, David S and Liu, Anqi}, |
| journal={arXiv preprint arXiv:2501.17871}, |
| year={2025} |
| } |
| ``` |
