中国科学院行为科学重点实验室知识库

当前精神障碍的诊断评估主要依赖于心理筛查量表和医生诊断,而很多精神障碍的临床表现比较复杂,且具有相似的外显临床症状,难以依靠现有测量方法从内在机制上进行精准区分.我们调研了国内外现有的心理实验测试任务集,分析并比较现有任务集的特征及优劣势,发现基于认知图谱构建心理实验标准化测试任务库,搭建网络和移动多种施测平台,有助于采集大规模的心理特征结构化数据.这些任务可用于构建客观、高效、可靠的精神障碍量化评估工具、指标和模型,以期对多重精神障碍进行多维度心理特征的量化评估,应用于复杂精神障碍进行精准分类,服务于精准医疗.更重要的是,标准化测试平台也将有助于正常人群认知行为测量,增强心理测试的标准性和可比性.

The current diagnostic assessment of mental illness mainly relies on psychological scales and the subjective judgment of doctors, and the clinical manifestations of many types of mental disorders are complex but very similar in the clinical symptoms, which cannot be accurately distinguished from the internal mechanism by existing measurement methods. At present, big data analysis has penetrated into various fields including medical diagnosis. These techniques rely on large amounts of structured annotated data and are currently lacking in psychometric testing. In recent years, the National Institutes of Health has proposed a diagnostic framework for mental illness based on Research Domain Criteria (RDoC), which attempts to differentiate mental disorders by internal mechanisms based on a large number of observable behavioral and neuroscientific indicators. The framework of the Cognitive Atlas is embodied in the RDoC. Cognitive Atlas is a tool for constructing and displaying interconnections between mental concepts and mental processes, mental states, and testing tasks. The standardized testing platform of psychological experiments based on Cognitive Atlas can break through the barriers of scattered data collection of psychological experiment task, single method and inability to collect structured data in large scale. In addition, it plays an important role in promoting the communication and comparison of psychological research results, improving the repeatability of research, understanding the national mental health level, and further promoting the construction of mental health. We investigated the task batteries that applied at home and abroad, and found that the representative task platforms abroad, such as CANTAB, NCANDA, and NIH Toolbox, were mainly developed for different mental diseases, while there are some traditional task testing platforms in China, mainly embodied in teaching, scientific research and the market application of mental health. We analyzed and compared the characteristics, advantages and disadvantages of the existing set of tasks, and concluded that foreign task platforms are mostly based on Cognitive Atlas construction, but the task coverage is not extensive enough, and there is no complete concept-task system. Most domestic test platforms only focus on some part of the test (such as questionnaires or cognitive tasks). Although the experimental paradigm on some test platforms is relatively comprehensive, it cannot meet the requirements of crossplatform large-scale testing, but mainly focuses on teaching mode. Moreover, the limitation of unity at home and abroad is that these platforms only focus on the part they are interested in, which is only the prototype of the Cognitive Atlas platform. Therefore, the establishment of a standardized experimental testing platform based on Cognitive Atlas and multiple terminals in China will meet the needs of improving the psychological experiment system and carrying out largescale and efficient structured data of psychological characteristics collection. Combined with these tasks and big data, we can construct the objective, efficient and reliable quantitative assessment tools, indicators and models for mental diseases, so as to carry out quantitative evaluation of multi-dimensional psychological characteristics of multiple mental disorders, apply them to accurate classification of complex mental diseases and serve precise medicine. More importantly, the standardized test platform will also contribute to the measurement of cognitive behavior of normal people and enhance the standardization and comparability of psychological tests. cognitive behavior, mental disorder, task platform, psychological experiment

北京市自然科学基金(5184035);; 国家重点研发计划(2016YFC0800900)资助