Background: This paper introduces a new clinical test, the Mini Linguistic State Examination (MLSE), as a short assessment for screening and classification of the different manifestations of primary progressive aphasia (PPA). Differentiation and monitoring of PPA variants are vital for management, planning and development of new treatments. The MLSE is designed to improve the uniformity of testing, screening for recruitment to clinical trials, and consistency of research results. It is a brief but effective test which can be adapted to the worlds major languages. Methods: Fifty-four patients and 30 age-, sex- and education-matched controls completed testing with the MLSE and components of the Boston Diagnostic Aphasia Examination in addition to their standard clinical diagnostic assessment. The MLSE includes five domains (motor speech, phonology, semantics, syntax and working memory) that were compared across groups. A random forest classification was used to learn the relationship between these five domains and assess the power of the diagnostic accuracy for predicting PPA subtypes. The final machine learning model was used to create a decision tree to guide the optimal manual classification of patients. Results: On average, the test took less than 20 minutes to administer. Significant group differences were found across all five domains, in terms of the distributions of error-types. These differences mirror the well-known language profiles for the three main PPA variants, which typically require an extended neuropsychology and speech pathology assessment. The random forest prediction model had an overall classification accuracy of 96% (92% for logopenic variant PPA, 93% for semantic variant PPA and 98% for non-fluent variant PPA). The derived decision tree for manual classification produced correct classification of 91% of participants whose data were not included in the training set. Conclusions: The MLSE is a new short cognitive test, with a scoring system that is easy to learn and apply. It is accurate for classifying PPA syndromes, and has potential to screen and monitor language deficits that occur in other focal and neurodegenerative brain disorders associated with language impairment. With increasing importance of language assessment in clinical research, the MLSEs linguistic assessment tool enables the essential profiling of language deficits in a wide clinical community.

中文翻译：

---

迷你语言状态考试（MLSE）：简要但准确的评估工具，用于对原发性失语症进行分类

背景：本文介绍了一种新的临床测试，即迷你语言状态检查（MLSE），作为对原发进行性失语症（PPA）不同表现形式的筛选和分类的简短评估。PPA变体的区分和监控对于新疗法的管理，计划和开发至关重要。MLSE旨在提高测试的一致性，筛查以招募至临床试验以及研究结果的一致性。这是一个简短而有效的测试，可以适应世界主要语言。方法：54名患者和30名年龄，性别和教育程度匹配的对照者除了进行标准的临床诊断评估外，还完成了MLSE和波士顿诊断失语症检查的组成部分的测试。MLSE包含五个领域（运动语音，语音，语义，语法和工作记忆）进行比较。使用随机森林分类来了解这五个域之间的关系，并评估诊断准确性对预测PPA亚型的影响。最终的机器学习模型用于创建决策树，以指导患者的最佳手动分类。结果：平均而言，该测试的执行时间不到20分钟。就错误类型的分布而言，在所有五个域中都发现了显着的组差异。这些差异反映了三种主要PPA变体的众所周知的语言配置文件，这通常需要扩展的神经心理学和言语病理学评估。随机森林预测模型的总体分类准确度为96％（对于隐身变体PPA，92％，语义变体PPA为93％，非流利变体PPA为98％。派生的用于人工分类的决策树对91％的参与者的正确分类进行了分类，这些参与者的数据未包含在训练集中。结论：MLSE是一种新的简短的认知测验，其评分系统易于学习和应用。它可以准确地对PPA综合征进行分类，并且具有筛查和监视与语言障碍相关的其他局灶性和神经退行性脑部疾病中出现的语言缺陷的潜力。随着语言评估在临床研究中的重要性日益提高，MLSEs语言评估工具可以在广泛的临床社区中对语言缺陷进行必要的剖析。派生的用于手动分类的决策树对91％的参与者的正确分类进行了分类，这些参与者的数据未包含在训练集中。结论：MLSE是一种新的短期认知测验，其评分系统易于学习和应用。它可以准确地对PPA综合征进行分类，并且具有筛查和监视与语言障碍相关的其他局灶性和神经退行性脑部疾病中出现的语言缺陷的潜力。随着语言评估在临床研究中的重要性日益提高，MLSEs语言评估工具可以在广泛的临床社区中对语言缺陷进行必要的剖析。派生的用于手动分类的决策树对91％的参与者的正确分类进行了分类，这些参与者的数据未包含在训练集中。结论：MLSE是一种新的短期认知测验，其评分系统易于学习和应用。它可以准确地对PPA综合征进行分类，并且具有筛查和监视与语言障碍相关的其他局灶性和神经退行性脑部疾病中出现的语言缺陷的潜力。随着语言评估在临床研究中的重要性日益提高，MLSEs语言评估工具可以在广泛的临床社区中对语言缺陷进行必要的剖析。它可以准确地对PPA综合征进行分类，并且具有筛查和监视与语言障碍相关的其他局灶性和神经退行性脑部疾病中出现的语言缺陷的潜力。随着语言评估在临床研究中的重要性日益提高，MLSEs语言评估工具可以在广泛的临床社区中对语言缺陷进行必要的剖析。它可以准确地对PPA综合征进行分类，并且具有筛查和监视与语言障碍相关的其他局灶性和神经退行性脑部疾病中出现的语言缺陷的潜力。随着语言评估在临床研究中的重要性日益提高，MLSEs语言评估工具可以在广泛的临床社区中对语言缺陷进行必要的剖析。