Construct Validation of the Verb Naming Test for Aphasia

Area of Science:

• Linguistics
• Neuroscience
• Psychology

Background:

• Spoken word production in aphasia is complex, with ongoing debate regarding specific verb production processes.
• Existing research suggests item properties like imageability may influence findings, but psychometric validation of verb tests is lacking.

Purpose of the Study:

• To investigate the cognitive constructs underlying the Verb Naming Test (VNT) using an item response theory (IRT) framework.
• To examine the predictive power of item (argument structure, imageability) and person (aphasia subtype, severity) covariates on VNT performance.

Main Methods:

• Utilized an archival dataset of 107 individuals with aphasia.
• Applied a series of IRT models to analyze VNT item response patterns.
• Assessed the influence of argument structure, imageability, aphasia subtype, and severity.

Main Results:

• Lexical-semantic processing covariates, specifically imageability and aphasia severity, were significant predictors of VNT performance.
• Morphosyntactic processing covariates, including argument structure and aphasia subtype, showed minimal predictive power.
• Imageability emerged as a key factor influencing VNT item responses.

Conclusions:

• VNT item response patterns are primarily driven by lexical-semantic factors, particularly imageability.
• The findings underscore the difficulty in distinguishing morphosyntactic from lexical-semantic processes in word production.
• Highlights the need to consider imageability in the design and interpretation of verb production tests for aphasia.