In recent years, applying machine learning methods to neurological and psychiatric disorder diagnoses has grasped the interest of many researchers; however, currently available machine learning toolboxes usually require somewhat intermediate programming knowledge. In order to use machine learning methods more quickly and conveniently, we developed an intuitive toolbox named BrainSort. BrainSort used Python as the main programming languages and employed a hospitable Graphical User Interface (GUI). The toolbox is user-friendly for researchers and clinical doctors with little to no prior programming skills. It enables the client to choose from multiple machine learning methods, such as support vector machine (SVM), k-nearest neighbors (k-NN), and convolutional neural network (CNN) for data processing and training. Using BrainSort, doctors and researchers can calculate and visualize the correlation between brain connectome topology parameters and the symptom in question without prolonged programming training, empowering them to use the powerful tool of machine learning in their studies and practices.

近年来，将机器学习方法应用于神经和精神疾病诊断已经引起了许多研究者的兴趣。但是，当前可用的机器学习工具箱通常需要一定程度的中级编程知识。为了更快，更方便地使用机器学习方法，我们开发了一个名为BrainSort的直观工具箱。BrainSort使用Python作为主要的编程语言，并采用了热情好客的图形用户界面（GUI）。该工具箱对于研究人员和临床医生来说非常友好，几乎没有或根本没有编程技能。它使客户端可以从多种机器学习方法中进行选择，例如支持向量机（SVM），k个最近邻居（k-NN），以及用于数据处理和训练的卷积神经网络（CNN）。使用BrainSort，医生和研究人员无需长时间的编程培训就可以计算和可视化大脑连接组拓扑参数与相关症状之间的相关性，从而使他们能够在研究和实践中使用强大的机器学习工具。