Acute post-stroke cognitive test performance predicts one-year functional independence

General cognitive ability (g), a latent variable derived from cognitive data, can predict life outcomes (e.g., educational attainment and occupational success) among neurologically healthy individuals. The value of g for predicting post-stroke functional outcomes is unknown. We addressed this gap here.

We derived g using exploratory structural equation modeling of 15 neuropsychological tests administered to 112 patients with stroke, 69 of whom also had functional outcome data (42 men; mean age = 53.23 years (SD = 10.54)). We used logistic regressions to compare g and individual tests in terms of their ability to predict the Functional Assessment Measure (FAM) and the Functional Independence Measure (FIM; motor and cognitive subscales) at 12 months post-stroke.

g was a statistically significant predictor of FAM (X ² = 6.86, p = 0.013) and the cognitive (X ² = 11.48, p = 0.002) but not motor FIM subscale (X ² = 0.93, p = 0.154). Individual tests varied widely in terms of predictive utility (X ² range: 0.07-21.03), with the most robust predictors being measures of visuospatial functions.

Acute measures of cognition, including g, can predict functional independence 12 months after stroke. g and visuospatial ability measures were the most robust predictors.