Auditory Verb Generation Performance Patterns Dissociate Variants of Primary Progressive Aphasia.

Area of Science:

• Neuroscience
• Linguistics
• Cognitive Psychology

Background:

• Primary progressive aphasia (PPA) is a neurodegenerative syndrome causing progressive language decline.
• PPA presents in three main variants: logopenic (lvPPA), semantic (svPPA), and non-fluent (nfvPPA), each linked to distinct brain regions and deficits.
• Differentiating PPA variants is crucial for accurate diagnosis and understanding underlying neuropathology.

Purpose of the Study:

• To investigate if distinct error patterns in a single language task can differentiate the three PPA variants.
• To explore the neurocognitive mechanisms underlying language deficits in each PPA subtype through error analysis.
• To assess the potential of error pattern analysis for enhancing PPA phenotyping and diagnostic sensitivity.

Main Methods:

• Recruited 21 lvPPA, 28 svPPA, 24 nfvPPA patients, and 31 healthy controls.
• Administered an auditory noun-to-verb generation task assessing auditory processing, lexical access, semantic retrieval, and speech production.
• Utilized machine learning to classify patients into PPA variants based on error types; analyzed specific error patterns.

Main Results:

• Task accuracy varied significantly across PPA variants and controls, with lvPPA showing lowest and nfvPPA highest accuracy.
• Machine learning successfully classified patients into their respective PPA variants based on error patterns.
• lvPPA patients produced more 'not-a-verb' responses and semantically related nouns; svPPA patients produced more 'unrelated verb' responses and light verbs.

Conclusions:

• Error patterns in an auditory verb generation task are specific to PPA variants, reflecting distinct neurocognitive deficits.
• Findings support the link between temporo-parietal regions and lexical processing (lvPPA), and anterior temporal lobe with semantic processes (svPPA).
• Analysis of response patterns offers valuable insights into PPA phenotyping, diagnostic sensitivity, and the neural basis of language components.