Shared Variance Across Cognitive Assessments Predicts Language Outcomes in Aphasia: A Preliminary Study

Aphasia, a common consequence of stroke, often co-occurs with cognitive impairments, yet clinical assessment is limited by tools that are lengthy, linguistically demanding, and sometimes redundant. This preliminary study examined shared variance across widely used cognitive tests and explored how the resulting principal components related to aphasia severity and sentence comprehension efficiency. We also tested whether single tasks, those with the strongest component loadings, could predict language outcomes as effectively as the full component scores.

Seventeen adults with chronic aphasia and 21 matched controls completed the Test of Everyday Attention, the Cognitive Linguistic Quick Test-Plus, and the Visual and Auditory Duration Attention Network Tests. Principal component analysis (PCA) was used to identify shared variance across cognitive tasks, followed by regression analyses relating components and single-task predictors to language outcomes.

PCA revealed six components, with subtests from the same battery generally clustering together, suggesting redundancy. Regression analyses showed that visuospatial processing speed and auditory orienting attention predicted less severe aphasia and more efficient sentence comprehension, whereas greater visual-sustained attention was linked to more severe aphasia. Single-task loadings predicted sentence comprehension but not aphasia severity.

These preliminary findings highlight the potential to streamline cognitive assessments to reduce redundancy, improve diagnostic precision, and better support individualized rehabilitation.

https://doi.org/10.23641/asha.33915754.