Data_Sheet_1_Visualizing Psychological Networks: A Tutorial in R.PDF

Networks have emerged as a popular method for studying mental disorders. Psychopathology networks consist of aspects (e.g., symptoms) of mental disorders (nodes) and the connections between those aspects (edges). Unfortunately, the visual presentation of networks can occasionally be misleading. For instance, researchers may be tempted to conclude that nodes that appear close together are highly related, and that nodes that are far apart are less related. Yet this is not always the case. In networks plotted with force-directed algorithms, the most popular approach, the spatial arrangement of nodes is not easily interpretable. However, other plotting approaches can render node positioning interpretable. We provide a brief tutorial on several methods including multidimensional scaling, principal components plotting, and eigenmodel networks. We compare the strengths and weaknesses of each method, noting how to properly interpret each type of plotting approach.

网络分析法已逐渐成为精神障碍研究领域的热门分析手段。心理病理网络以精神障碍的各类维度（如症状）作为节点（nodes），并将这些维度之间的关联作为边（edges）。然而，网络的可视化呈现有时会带来误导性解读。例如，研究者容易误以为，空间上彼此邻近的节点关联性较强，而间距较远的节点关联性较弱，但实际情况并非总是如此。作为目前最主流的可视化方法，力导向算法（force-directed algorithms）所绘制的网络中，节点的空间排布难以直观解读。不过，其他可视化方法可使节点位置的排布具备可解释性。本研究提供了涵盖多维标度法（multidimensional scaling）、主成分绘图法（principal components plotting）以及特征模型网络（eigenmodel networks）在内的多种方法的简要教程，并对比了各方法的优劣，同时阐明了如何正确解读各类可视化绘图方法的结果。