Comparing short and long batteries to assess deficits and their neural bases in stroke aphasia

Multiple language assessments are necessary for diagnosing, characterising and quantifying the multifaceted deficits observed in many patients post-stroke. Current language batteries, however, tend to be an imperfect trade-off between time and sensitivity of assessment. There have hitherto been two main types of battery. Extensive batteries provide thorough information but are impractically long for application in clinical settings or large-scale research studies. Clinically-targeted batteries tend to provide superficial information about a large number of language skills in a relatively short period of time by reducing the depth of each test but, consequently, can struggle to identify mild deficits, qualify the level of each impairment or reveal the underlying component structure. In the current study, we compared these batteries across a large group of individuals with chronic stroke aphasia to determine their utility. In addition, we developed a data-driven reduced version of an extensive battery that maintained sensitivity to mild impairment, ability to grade deficits and the component structure. The underlying structure of these three language batteries (extensive, shallow and data-reduced) was analysed using cross-validation analysis and principal component analysis. This revealed a four-factor solution for the extensive and data-reduced batteries, identifying phonology, semantic skills, fluency and executive function in contrast to a two-factor solution using the shallow battery (phonological/language severity and cognitive severity). Lesion symptom mapping using participants factor scores identified convergent neural structures based on existing language models for phonology (superior temporal gyrus), semantics (inferior temporal gyrus), speech fluency (precentral gyrus) and executive function (lateral occipitotemporal cortex) based on the extensive and data-reduced batteries. The two components in the shallow battery converged with the phonology and executive function clusters. In addition, we show that multivariate prediction models could be utilised to predict the component scores using neural data, however not for every component score within every test battery. Overall, the data-reduced battery appears to be an effective way to save assessment time yet retain the underlying structure of language and cognitive deficits observed in post stroke aphasia.