Measuring word retrieval in narrative discourse: core lexicon in aphasia

Abstract

Background Discourse analysis procedures are time consuming and impractical in a clinical setting. Critical to clinicians are simple and informative discourse measures that require minimal time and labour to complete. Many studies, however, have overlooked difficulties that clinicians face. We recently developed core lexicon lists for nouns, verbs, adjectives and adverbs for two narrative discourse tasks with healthy control groups. Core lexicon lists consist of important lexical items required to produce coherently meaningful discourse in response to discourse tasks. Measuring core lexicon is useful for quantifying word retrieval impairments at the discourse level in clinical populations. Aims To apply an age‐based core lexicon list for nouns, verbs, adjectives and adverbs for the wordless picture books Good Dog Carl (1985) and Picnic (1984) and to determine how well the lists measured linguistic impairments in persons with aphasia (PWA). Materials & Methods Lemma forms were extracted from 470 control participants who were divided into seven age groups. Twenty‐five core lexicons were identified for four word classes (nouns, verbs, adjectives and adverbs) among the seven age groups. The nouns, verbs, adjectives and adverbs for each PWA (N = 11) were then compared with the core lexicon for their respective age group. Per cent agreement was computed by comparing the number of total items within each list to the number of items that PWA produced. A Spearman's correlation coefficient was computed between the WAB‐R AQ and the per cent agreement for each word type for PWA. Outcomes & Results The percentage of agreement for each word type among the age cohorts ranged between 56% and 96%. Of the four word types, core verbs significantly correlated with the WAB AQs for both discourse tasks. A post‐hoc analysis found significant differences between fluent and non‐fluent aphasia for core verbs. Conclusions & Implications Core lexicon analysis appears to be a practical way to capture impairments in word retrieval at the discourse level. Core verbs may be a better indicator to understand holistic language performances for PWA. Use of the core lexicon checklist can serve as an option to reconcile ecological validity with clinical usability.

A Compendium of Core Lexicon Checklists

“…7,8,15 Wordless Picture Book Tasks Kim and colleagues developed CoreLex checklists for the wordless picture books "Good Dog Carl" and "Picnic." 9 Checklists were developed from samples collected from 470 healthy control speakers, relatively evenly distributed across seven age bins (20s, 30s, 40s, 50s, 60s, 70s, and 80s) and approximately matched for sex and education. Separate checklists were created for four word types (nouns, verbs, adjectives, and adverbs), given conflicting evidence that PWAs experience differential impairment of production of word types during discourse.…”

“…4 Indeed, research examining CoreLex indicates that PWAs produce fewer typical lexical items than controls. [5][6][7][8] CoreLex has demonstrated sensitivity to age-related differences in neurologically healthy individuals, 9 between healthy individuals and PWAs, and between aphasia subtypes. 7,8 CoreLex also correlated with performance on standardized assessments and other discourse measures.…”

“…7,8 CoreLex also correlated with performance on standardized assessments and other discourse measures. [7][8][9][10][11] For a discussion of additional potential applications of CoreLex, please see the article by Dalton et al 12 CoreLex checklists have been reported for a variety of discourse tasks. The majority have been developed using stimuli from the Aphasia-Bank database (aphasia.talkbank.org).…”

“…13 In addition, CoreLex checklists were recently developed for two wordless picture books (Good Dog Carl, Picnic) classified as story generation tasks, as the stimuli are more complex (like Cinderella), but should be unfamiliar to most (unlike Cinderella). 9 Detailed elicitation instructions for these tasks are available in the article by Kim and Wright. 4 Unfortunately, these checklists are dispersed throughout the literature 5,8,9,14 or only available as conference proceedings.…”

“…9 Detailed elicitation instructions for these tasks are available in the article by Kim and Wright. 4 Unfortunately, these checklists are dispersed throughout the literature 5,8,9,14 or only available as conference proceedings. 6,7,15 Furthermore, CoreLex checklists have been developed by different groups and for different purposes, which means that the results of each checklist may need to be interpreted differently and/or different checklists should be used for different purposes.…”