Deep neural networks are increasingly used for neurological disease classification by MRI, but the networks’ decisions are not easily interpretable by humans. Heat mapping by deep Taylor decomposition revealed that (potentially misleading) image features even outside of the brain tissue are crucial for the classifier’s decision. We propose a regularization technique to train convolutional neural network (CNN) classifiers utilizing relevance-guided heat maps calculated online during training. The method was applied using T1-weighted MR images from 128 subjects with Alzheimer’s disease (mean age=71.9±8.5 years) and 290 control subjects (mean age=71.3±6.4 years). The developed relevance-guided framework achieves higher classification accuracies than conventional CNNs but more importantly, it relies on less but more relevant and physiological plausible voxels within brain tissue. Additionally, preprocessing effects from skull stripping and registration are mitigated. With the interpretability of the decision mechanisms underlying CNNs, these results challenge the notion that unprocessed T1-weighted brain MR images in standard CNNs yield higher classification accuracy in Alzheimer’s disease than solely atrophy.

中文翻译：

---

可解释的脑疾病分类和相关性引导的深度学习

深度神经网络越来越多地用于通过 MRI 对神经系统疾病进行分类，但这些网络的决定并不容易被人类解释。通过深度泰勒分解进行的热图显示，即使在脑组织之外（可能具有误导性）的图像特征对于分类器的决策也是至关重要的。我们提出了一种正则化技术，利用训练期间在线计算的相关引导热图来训练卷积神经网络 (CNN) 分类器。该方法使用来自 128 名阿尔茨海默病受试者（平均年龄 = 71.9±8.5 岁）和 290 名对照受试者（平均年龄 = 71.3±6.4 岁）的 T1 加权 MR 图像应用。开发的相关性引导框架比传统的 CNN 实现了更高的分类准确度，但更重要的是，它依赖于脑组织内更少但更相关和生理上合理的体素。此外，减轻了头骨剥离和配准的预处理效果。由于 CNN 的决策机制的可解释性，这些结果挑战了标准 CNN 中未处理的 T1 加权脑 MR 图像在阿尔茨海默病中产生比单纯萎缩更高分类准确度的观点。