Mauri Chiara, Cerri Stefano, Puonti Oula, Mühlau Mark, Van Leemput Koen
Department of Health Technology, Technical University of Denmark, Denmark; Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, USA.
Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, USA.
Med Image Anal. 2025 Apr;101:103436. doi: 10.1016/j.media.2024.103436. Epub 2024 Dec 27.
Recent years have seen a growing interest in methods for predicting an unknown variable of interest, such as a subject's diagnosis, from medical images depicting its anatomical-functional effects. Methods based on discriminative modeling excel at making accurate predictions, but are challenged in their ability to explain their decisions in anatomically meaningful terms. In this paper, we propose a simple technique for single-subject prediction that is inherently interpretable. It augments the generative models used in classical human brain mapping techniques, in which the underlying cause-effect relations can be encoded, with a multivariate noise model that captures dominant spatial correlations. Experiments demonstrate that the resulting model can be efficiently inverted to make accurate subject-level predictions, while at the same time offering intuitive visual explanations of its inner workings. The method is easy to use: training is fast for typical training set sizes, and only a single hyperparameter needs to be set by the user. Our code is available at https://github.com/chiara-mauri/Interpretable-subject-level-prediction.
近年来，人们对从描绘解剖功能效应的医学图像中预测未知感兴趣变量（如受试者的诊断）的方法越来越感兴趣。基于判别建模的方法在进行准确预测方面表现出色，但在以解剖学上有意义的术语解释其决策能力方面面临挑战。在本文中，我们提出了一种本质上可解释的单受试者预测简单技术。它用一个捕捉主导空间相关性的多变量噪声模型增强了经典人类脑图谱技术中使用的生成模型，在该生成模型中可以编码潜在的因果关系。实验表明，所得模型可以有效地求逆以进行准确的受试者水平预测，同时对其内部工作原理提供直观的视觉解释。该方法易于使用：对于典型的训练集大小，训练速度很快，并且用户只需设置一个超参数。我们的代码可在https://github.com/chiara - mauri/Interpretable - subject - level - prediction获取。
