Automated analysis of language production in aphasia and right-hemisphere damage: frequency and collocation strength

Abstract

ABSTRACTBackground: Reliance on formulaic language, i.e., holistically processed multiword chunks, is claimed to distinguish speakers with aphasia, speakers with right-hemisphere damage (RHD), and neurotypical controls (NC). Frequency and collocation strength of word combinations are indicators of formulaic language.Aims: We aimed to determine frequency and collocation properties of spoken production, as indicators of formulaic language, in people with aphasia, speakers with RHD and NC.Methods and Procedures: We used computerized methods to investigate spontaneous language in 40 speakers: 10 with fluent aphasia (fA), 10 with non-fluent aphasia (nfA), 10 with RHD, and 10 NC. Our analysis not only focused on frequency and collocation strength of grammatical combinations as markers of formulaic language (using the British National Corpus as reference), but also looked at word frequency, language fluency, proportion of content words, and measures of lexical and combinatorial diversity.Outcomes and Results: Both groups with aphasia differed from neurotypical speakers with regard to lexical features and word combinations. Their language was less fluent, less diverse at the word level, and, in the non-fluent group, contained a higher proportion of content words. Each aphasic group also differed significantly from controls with increased values on at least one marker of formulaic language. Speakers with RHD produced less-frequent combinations, which were more weakly collocated; however, these effects did not reach statistical significance.Conclusions: Our results show that formulaic language use distinguishes aphasic from other speakers and that differences can be tracked with an automated, corpus-based script that uses frequency and collocation variables. We present our study in the context of usage-based frameworks of language.