Prevalence of Verb and Sentence Impairment in Aphasia as Demonstrated by Cluster Analysis

Abstract

ABSTRACT Background: Verb and sentence processing impairments are frequently observed in people with aphasia (PWA). However, patterns of impairment (i.e. combinations of deficits that are revealed by arrays of tasks) are usually investigated at the group level, by selecting groups of patients with specific aphasia types (e.g., non-fluent variants with signs of agrammatism), or in case-series with relatively small samples. Aims: We aimed to explore patterns of impairment in verb and sentence processing in a large (N = 54) group of PWA with various aphasia types, using methods that do not require a priori aphasia-type information. Methods & Procedures: To assess verb and sentence processing impairments in production and comprehension, we used the Verb and Sentence Test (VAST) adapted to Russian (VAST-Ru). As our main method, we used k-means cluster analysis to group individuals based on their scores on different tasks of the VAST-Ru. We then compared cluster allocations to aphasia types (according to Alexander Luria’s classification) established by a standard clinical evaluation procedure. As a follow up, we investigated performance of each person with aphasia on each task by comparing their individual scores to the control group. Outcomes & Results: The k-means analysis grouped the PWA into two clusters. The main difference between the clusters was the severity of impairment across most of the tasks, both at the single verb and at the sentence level, both in production and in comprehension. All participants with efferent motor aphasia (roughly corresponding to Broca’s aphasia) and most of the participants with acoustic-mnestic aphasia (roughly corresponding to anomic aphasia) were in the less-impaired cluster. Complex aphasia types (complex motor and complex temporal aphasia) pertained to the more-impaired cluster, while other aphasia types (dynamic aphasia, corresponding to transcortical motor, and sensory, corresponding to Wernicke’s aphasia) were more or less evenly distributed over two clusters. Individual-level analysis revealed heterogeneous patterns of impairment in both clusters and across different aphasia types. Conclusions: These results demonstrate that verb and sentence impairments are prevalent in aphasia, irrespective of the aphasia type. They also indicate that aphasia severity mediates verb and sentence impairment in aphasia. We could not find a clear correspondence between individual patterns of impairment and aphasia type at this stage. The clinical implications are discussed.