Interpretable and integrative deep learning for discovering brain-behaviour associations

This record contains the data of the paper "Interpretable and integrative deep learning for discovering brain-behaviour associations". In this paper, we employ a digital avatar procedure as an interpretability module capable of reporting the relationships learned within a multi-view variational autoencoder. We integrate this procedure into a novel framework that utilises stability selection to identify meaningful and reproducible associations between brain imaging and behaviour.本数据集包含论文《可解释与整合式深度学习用于发现脑-行为关联》的相关数据。在该研究中，我们采用数字化身（digital avatar）流程作为可解释性模块，能够对多视图变分自编码器（multi-view variational autoencoder）内部学习得到的关联关系进行阐释。我们将该流程整合至一款新颖框架中，该框架借助稳定性选择（stability selection）方法，以识别脑成像（brain imaging）与行为之间具有意义且可复现的关联。