ABSTRACT Background The concept of fluency is widely used to dichotomously classify aphasia syndromes in both research and clinical practice. Despite its ubiquity, reliability of fluency measurement is reduced due to its multi-dimensional nature and the variety of methods used to measure it. Aims The primary aim of the study was to determine what factors contribute to judgements of fluency in aphasia, identifying methodological and linguistic sources of disagreement. Methods & Procedures We compared fluency classifications generated according to fluency scores on the revised Western Aphasia Battery (WAB-R) to clinical impressions of fluency for 254 English-speaking people with aphasia (PwA) from the AphasiaBank database. To determine what contributed to fluency classifications, we examined syndrome diagnoses and measured the predictive strength of 18 spontaneous speech variables extracted from retellings of the Cinderella story. The variables were selected to represent three dimensions predicted to underlie fluency: grammatical competence, lexical retrieval, and the facility of speech production. Outcomes & Results WAB-R fluency classifications agreed with 83% of clinician classifications, although agreement was much greater for fluent than nonfluent classifications. The majority of mismatches were diagnosed with anomic or conduction aphasia by the WAB-R but Broca’s aphasia by clinicians. Modifying the WAB-R scale improved the extent to which WAB-R fluency categories matched clinical impressions. Fluency classifications were predicted by a combination of variables, including aspects of grammaticality, lexical retrieval and speech production. However, fluency classification by WAB-R was largely predicted by severity, whereas the presence or absence of apraxia of speech was the largest predictor of fluency classifications by clinicians. Conclusions Fluency judgements according to WAB-R scoring and those according to clinical impression showed some common influences, but also some differences that contributed to mismatches in fluency categorization. We propose that, rather than using dichotomous fluency categories, which can mask sources of disagreement, fluency should be explicitly identified relative to the underlying deficits (word-finding, grammatical formulation, speech production, or a combination) contributing to each individual PwA’s fluency profile. Identifying what contributes to fluency disruptions is likely to generate more reliable diagnoses and provide more concrete guidance regarding therapy, avenues we are pursuing in ongoing research.

中文翻译：

---

流利还是不流利？A 部分. 失语症流畅度分类的潜在贡献者

摘要 背景 在研究和临床实践中，流畅性的概念被广泛用于对失语症综合征进行二分类。尽管它无处不在，但由于其多维性质和用于测量它的各种方法，流畅度测量的可靠性降低了。目的 本研究的主要目的是确定哪些因素有助于判断失语症的流畅性，确定方法和语言上的分歧来源。方法和程序 我们将根据修订后的西方失语症电池 (WAB-R) 的流畅度评分生成的流畅度分类与来自 AphasiaBank 数据库的 254 名讲英语的失语症患者 (PwA) 的流畅度临床印象进行了比较。为了确定是什么促成了流畅度分类，我们检查了综合征诊断并测量了从灰姑娘故事复述中提取的 18 个自发语音变量的预测强度。选择变量来代表预测流利基础的三个维度：语法能力、词汇检索和言语产生的能力。结果与结果 WAB-R 流畅度分类与 83% 的临床医生分类一致，尽管流畅度分类的一致性比非流畅分类大得多。大多数不匹配被 WAB-R 诊断为失语或传导性失语，但临床医生诊断为 Broca 失语。修改 WAB-R 量表提高了 WAB-R 流畅度类别与临床印象的匹配程度。流利度分类是通过变量的组合来预测的，包括语法、词汇检索和语音生成。然而，WAB-R 的流畅度分类主要是根据严重程度来预测的，而言语失用症的存在与否是临床医生对流畅度分类的最大预测因素。结论 WAB-R评分和临床印象的流畅度判断显示出一些共同的影响，但也有一些差异导致了流畅度分类的不匹配。我们建议，与其使用可以掩盖分歧来源的二分流利度类别，不如明确确定流利度相对于导致每个 PwA 流利度概况的潜在缺陷（找词、语法表述、语音生成或组合） .