This is the code for "Finding Essential Parts of the Brain in rs-fMRI can Improve ADHD Diagnosis using Deep Learning" in Arxiv (https://arxiv.org/abs/2108.10137)

This repository has only code related with paper. If you want to run this code, you need to prepare a few thing.

• Tensorflow == 2.4
• scikit-learn >= 0.24.1
• matpotlib >= 3.3.4

This is not contain fMRI data. If you want to download the data, see http://preprocessed-connectomes-project.org/adhd200/download.html

You can find NIAK dataset in https://www.nitrc.org/frs/?group_id=383

If you download the data from section 1.2, you need to preprocessing the data. We preprocessed data using AAL 116 atlas with SPM12

• main.py : model training code with hyperparameter setting
• models.py : SCCNN type's models (e.g., SCCNN-RNN, ASCRNN, ASDRNN, ASSRNN)
• layers.py : SCCNN block and attention layer contained
• dataset.py : make dataset from preprcessed data
• result.py : summerize the results