Does error type predict underlying impairment in anomia?

Background: Aphasic impairments of lexical retrieval (LR) can be attributed to different underlying deficits such as problems with meaning (semantic deficits), with word selection (lexical deficits) and sound selection (phonological deficits). Research in the area is unclear on the extent to which types of errors produced are a reliable indicator of the underlying impairment. It is also unclear in the literature whether recording of all errors made on a picture naming task can reveal more about LR breakdown than recording only first responses. Aims: To determine if analyses of error types produced by ‘first responses’ or ‘all responses’ given on a naming task provide more accurate information for differential diagnosis of impairment in spoken word production. Additionally, to establish whether error type predicts underlying impairment in anomia. Methods & Procedures: Cross-sectional data was obtained from 34 Irish participants who were diagnosed with aphasia secondary to a cerebrovascular accident. Naming performance was assessed by means of the complete set of the Boston Naming Test (BNT; Kaplan et al. 2001). Additional tasks of comprehension and repetition were also carried out using subtests of the Comprehensive Aphasia Battery (CAT; Swinburn et al. 2004). Participants were categorised into ‘main error type’ groups. Level of breakdown in spoken word production was predicted by results of comprehension and repetition tests. This was then compared with BNT error types produced. Outcomes & Results: Analyses of error profiles indicated that there was no significant difference between first responses given on the BNT and all responses given. Therefore, first responses provide an accurate account of an individual’s error profile. Chi-square analysis reported that there was a significant association between main error types produced and the predicted level of breakdown in LR. Conclusion: First responses given on a naming task may be sufficient in predicting the level of breakdown in LR. However, additional tasks of comprehension and repetition may be of clinical benefit to confirm a diagnosis, and to plan treatment accordingly.