Detection of ADHD based on Eye Movements during Natural Viewing

In this paper, we explore whether Attention-deficit/hyperactivity disorder (ADHD) can be detected based on recorded eye movements together with information about the video stimulus in a free-viewing task

Setup

Clone repository:

git clone git@github.com:aeye-lab/ecml-ADHD

or

git clone https://github.com/aeye-lab/ecml-ADHD

and change to the cloned repo via cd ecml-ADHD.

Install dependencies:

pip install -r requirements.txt

Run Experiments

Prepare saliency maps

Please download videos from http://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/EEG-Eyetracking%20Protocol.html, and break it down into frames (e.g. use ffmpeg packages) and put them under the folder /Data/videos/frames_{video_name}/

We use a state-of-the-art saliency model, DeepGazeII to compute salinecy maps for our video stimuli. Download the files you need to run the DeepGazeII model https://drive.google.com/file/d/1kYUwoatqQUS5EabeeSDc6gRmCysnVZ6N/view, and stored under the folder /DNN_model/DataGeneration/

To generate the saliency maps for all videos, run

bash gen_saliency_map_data.sh

Prepare model input files

To generate model input files, run

bash gen_model_input_data.sh

Run models

bash run_models.sh

Cite our work

If you use our code for your research, please consider citing our paper:

@inproceedings{deng2023detection,
  title={Detection of ADHD based on eye movements during natural viewing},
  author={Deng, Shuwen and Prasse, Paul and Reich, David R and Dziemian, Sabine and Stegenwallner-Sch{\"u}tz, Maja and Krakowczyk, Daniel and Makowski, Silvia and Langer, Nicolas and Scheffer, Tobias and J{\"a}ger, Lena A},
  booktitle={Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2022, Grenoble, France, September 19--23, 2022, Proceedings, Part VI},
  pages={403--418},
  year={2023},
  organization={Springer}
}