Objectives: Semantic feature analysis (SFA) is a treatment to improve word retrieval ability by strengthening impaired semantic networks. A large number of studies have reported the efficacy of SFA treatment on trained items, but its generalization effects on untrained items, discourse production or standardized language measures are controversial. Through the use of meta-analyses, the current study aimed to provide a systematic review of the treatment and generalization effects of SFA for individuals with aphasia. Methods: A systematic search based on 5 databases (DBPIA, RISS, EBSCOhost, ProQuest, PubMed) identified 11 studies which met the inclusion criteria. Effect sizes were calculated using Cohen's d and robust improvement rate difference for naming and discourse production. We analyzed the standardized mean difference for standardized language measures. Results: Results demonstrated medium to large effect sizes for trained items, but small effect sizes for untrained items. For discourse production, effect sizes varied from small to large depending on the types of outcome measures. SFA treatment approach contributed to improving overall language ability, but its generalization effects onto naming domains from standardized language tests seem to be limited. Conclusion: SFA treatment was effective for improving naming ability on trained items and increasing overall language ability. However, its generalization effects were relatively limited for untrained items, discourse production and standardized naming tests. These results suggest that researchers and clinicians should consider several factors which may affect the treatment efficacy of a SFA approach for individuals with aphasia.
- Semantic feature analysis