Charité - Universitätsmedizin Berlin, Corporate Member Of Freie Universität Berlin, Humboldt-Universität zu Berlin; Department of Psychiatry and Psychotherapy, 10117 Berlin, Germany; Bernstein Center for Computational Neuroscience, 10117 Berlin, Germany.
Charité - Universitätsmedizin Berlin, Corporate Member Of Freie Universität Berlin, Humboldt-Universität zu Berlin; Department of Psychiatry and Psychotherapy, 10117 Berlin, Germany; Bernstein Center for Computational Neuroscience, 10117 Berlin, Germany.
Exp Neurol. 2021 May;339:113608. doi: 10.1016/j.expneurol.2021.113608. Epub 2021 Jan 26.
By promising more accurate diagnostics and individual treatment recommendations, deep neural networks and in particular convolutional neural networks have advanced to a powerful tool in medical imaging. Here, we first give an introduction into methodological key concepts and resulting methodological promises including representation and transfer learning, as well as modelling domain-specific priors. After reviewing recent applications within neuroimaging-based psychiatric research, such as the diagnosis of psychiatric diseases, delineation of disease subtypes, normative modeling, and the development of neuroimaging biomarkers, we discuss current challenges. This includes for example the difficulty of training models on small, heterogeneous and biased data sets, the lack of validity of clinical labels, algorithmic bias, and the influence of confounding variables.
通过承诺更准确的诊断和个性化的治疗建议，深度神经网络，特别是卷积神经网络已经成为医学成像领域的有力工具。在这里，我们首先介绍方法学的关键概念和由此产生的方法学承诺，包括表示和迁移学习，以及针对特定领域的先验建模。在回顾了基于神经影像学的精神疾病研究中的最新应用，如精神疾病的诊断、疾病亚型的划定、规范建模以及神经影像学生物标志物的开发之后，我们讨论了当前的挑战。例如，在小、异质和有偏差的数据集上训练模型的困难、临床标签的有效性缺失、算法偏差以及混杂变量的影响。
