Connected Speech Alterations and Progression in Patients With Primary Progressive Aphasia Variants.

Area of Science:

• Neurology
• Speech-Language Pathology
• Neuroimaging

Background:

• Diagnosing primary progressive aphasia (PPA) variants is challenging, impacting patient management.
• Accurate characterization of PPA subtypes is crucial for improving treatment outcomes.

Purpose of the Study:

• To identify speech features, alone or combined with language assessments and gray matter volumes (GMVs), that best distinguish PPA variants.
• To analyze the evolution of connected speech in different PPA variants over time.

Main Methods:

• Prospective study including 95 PPA patients with neuropsychological assessments, "Picnic Scene" speech test, and brain MRI.
• Sequential feature selection models incorporated speech parameters, language tests, and GMVs to differentiate PPA groups.
• Linear mixed-effect models analyzed speech changes over time within each PPA variant.

Main Results:

• Models combining speech parameters and GMVs significantly improved accuracy in differentiating nonfluent variant PPA (nfvPPA) and logopenic variant PPA (lvPPA) compared to standard language scores alone.
• Semantic variant PPA (svPPA) was accurately distinguished from other PPA groups by all models.
• Distinct speech trajectories were observed: nfvPPA showed more phonological errors, svPPA exhibited semantic/morphosyntactic errors, and lvPPA had reduced speech rate and sentence length.

Conclusions:

• Incorporating "Picnic Scene" speech test measures and GMVs enhances the accuracy of differentiating PPA variants, particularly nfvPPA and lvPPA.
• Speech features and GMVs are valuable tools for understanding PPA progression and predicting patient outcomes.
• Distinct speech evolution patterns in PPA variants can inform targeted speech therapy interventions.