Wehenkel Marie, Sutera Antonio, Bastin Christine, Geurts Pierre, Phillips Christophe
Department of Computer Science and Electrical Engineering, Montefiore Institute, University of Liège, Liège, Belgium.
GIGA-CRC in silico Medicine, University of Liège, Liège, Belgium.
Front Neurosci. 2018 Jun 29;12:411. doi: 10.3389/fnins.2018.00411. eCollection 2018.
Machine learning approaches have been increasingly used in the neuroimaging field for the design of computer-aided diagnosis systems. In this paper, we focus on the ability of these methods to provide interpretable information about the brain regions that are the most informative about the disease or condition of interest. In particular, we investigate the benefit of group-based, instead of voxel-based, analyses in the context of Random Forests. Assuming a prior division of the voxels into non overlapping groups (defined by an atlas), we propose several procedures to derive group importances from individual voxel importances derived from Random Forests models. We then adapt several permutation schemes to turn group importance scores into more interpretable statistical scores that allow to determine the truly relevant groups in the importance rankings. The good behaviour of these methods is first assessed on artificial datasets. Then, they are applied on our own dataset of FDG-PET scans to identify the brain regions involved in the prognosis of Alzheimer's disease.
机器学习方法在神经成像领域越来越多地用于计算机辅助诊断系统的设计。在本文中，我们关注这些方法提供有关对感兴趣的疾病或状况最具信息性的脑区的可解释信息的能力。特别是，我们研究了在随机森林的背景下基于组而不是基于体素的分析的益处。假设体素预先划分为非重叠组（由图谱定义），我们提出了几种程序，从随机森林模型得出的个体体素重要性中推导组重要性。然后，我们采用几种置换方案，将组重要性分数转换为更具可解释性的统计分数，以便在重要性排名中确定真正相关的组。这些方法的良好性能首先在人工数据集上进行评估。然后，将它们应用于我们自己的FDG-PET扫描数据集，以识别与阿尔茨海默病预后相关的脑区。
