Naming errors and dysfunctional tissue metrics predict language recovery after acute left hemisphere stroke

Area of Science:

• Neurology
• Neuroscience
• Speech-Language Pathology

Background:

• Predicting language recovery after left hemisphere (LH) stroke is difficult using global measures or lesion size alone.
• Existing methods for predicting post-stroke language recovery are insufficient.

Purpose of the Study:

• To investigate if naming error types and dysfunctional brain tissue metrics improve predictions of language recovery after acute LH stroke.
• To determine the relationship between specific naming error profiles and brain damage with language outcomes over time.

Main Methods:

• 148 individuals with acute LH stroke underwent neuroimaging and language assessments.
• Naming errors were categorized (semantic, phonological, mixed, unrelated) and analyzed with brain tissue damage and hypoperfusion metrics.
• Language skills were reassessed at subacute, early chronic (5-7 months), and late chronic (≥11 months) phases.

Main Results:

• Higher proportions of semantic errors correlated with better acute naming, while other error types indicated poorer accuracy.
• Language skills improved over time, but naming error profiles remained stable.
• Fewer unrelated errors and less damage to the left angular gyrus predicted optimal recovery.
• Greater acute errors and left middle temporal gyrus damage were linked to the largest language gains.

Conclusions:

• Naming error profiles, especially unrelated errors, significantly enhance the prediction of language recovery post-stroke.
• Specific error types and lesion locations provide distinct insights into recovery trajectories and potential for improvement.