Semantic feature analysis and aphasia (Gravier et al., 2018)

Purpose: This study investigated the predictive value of practice-related variables—number of treatment trials delivered, total treatment time, average number of trials per hour, and average number of participant-generated features per trial—in response to semantic feature analysis (SFA) treatment.Method: SFA was administered to 17 participants with chronic aphasia daily for 4 weeks. Individualized treatment and semantically related probe lists were generated from items that participants were unable to name consistently during baseline testing. Treatment was administered to each list sequentially in a multiple-baseline design. Naming accuracy for treated and untreated items was obtained at study entry, exit, and 1-month follow-up.Results: Item-level naming accuracy was analyzed using logistic mixed-effect regression models. The average number of features generated per trial positively predicted naming accuracy for both treated and untreated items, at exit and follow-up. In contrast, total treatment time and average trials per hour did not significantly predict treatment response. The predictive effect of number of treatment trials on naming accuracy trended toward significance at exit, although this relationship held for treated items only.Conclusions: These results suggest that the number of patient-generated features may be more strongly associated with SFA-related naming outcomes, particularly generalization and maintenance, than other practice-related variables.Supplemental Material S1. Supplemental language testing results.Supplemental Material S2. Determination of eligible treatment items based on naming task performance (sample).Supplemental Material S3. Eligible items by category (sample).Supplemental Material S4. Treatment fidelity checklist.Supplemental Material S5. Treatment list progression based on naming probe accuracy (sample). Gravier, M. L., Dickey, M. W., Hula, W. D., Evans, W. S., Owens, R. L., Winans-Mitrik, R. L., & Doyle, P. J. (2018). What matters in semantic feature analysis: Practice-related predictors of treatment response in aphasia. American Journal of Speech-Language Pathology, 27(1S), 438–453.

本研究旨在探讨与练习相关的变量（包括治疗试验次数、总治疗时间、每小时平均试验次数以及每次试验中参与者生成的特征平均数量）对语义特征分析（SFA）治疗的预测价值。研究方法为每日对17名慢性失语症患者进行SFA治疗，为期4周。根据基线测试中参与者无法持续命名的项目，生成个性化的治疗和语义相关探测列表。治疗以多基线设计依次对每个列表进行。在研究开始、结束及1个月随访时，记录治疗和未治疗项目的命名准确性。研究结果：采用逻辑混合效应回归模型分析了项目层面的命名准确性。每次试验生成的特征平均数量对治疗和未治疗项目的命名准确性均呈正相关，无论是在结束还是随访阶段。相反，总治疗时间和每小时平均试验次数对治疗反应没有显著预测作用。治疗试验次数对命名准确性的预测效应在结束阶段趋向于显著性，尽管这种关系仅适用于治疗项目。结论：这些结果表明，患者生成的特征数量可能与SFA相关的命名结果（尤其是泛化和维持）比其他练习相关变量有更强的关联性。补充材料S1：补充语言测试结果。补充材料S2：基于命名任务表现确定合格的治疗项目（样本）。补充材料S3：按类别划分的合格项目（样本）。补充材料S4：治疗一致性检查清单。补充材料S5：基于命名探测准确性的治疗列表进展（样本）。Gravier, M. L.，Dickey, M. W.，Hula, W. D.，Evans, W. S.，Owens, R. L.，Winans-Mitrik, R. L.，& Doyle, P. J.（2018）。语义特征分析中的关键因素：失语症治疗反应的练习相关预测因子。美国言语-语言病理学杂志，27（1S），438–453。