Biomarkers

Area of Science:

• Neuroscience
• Computational Linguistics
• Biomarker Discovery

Background:

• Alzheimer's disease (AD) presents diverse clinical forms requiring tailored care.
• Logopenic variant primary progressive aphasia (lvPPA) prognosis is challenging due to varied underlying pathologies (AD or FTLD).
• Accurate diagnosis necessitates multidisciplinary approaches including CSF analysis, imaging, and psychometric assessments.

Purpose of the Study:

• To evaluate speech markers for cross-classifying AD phenotypes with Cerebro-Spinal Fluid (CSF) profiles.
• To provide a non-invasive method for early prognosis of neurodegenerative evolution.
• To assess the potential of machine learning models in differentiating AD subtypes and pathologies.

Main Methods:

• 42 patients were classified into lvPPA- (FTLD), lvPPA+ (AD), and AD groups.
• Speech recordings from a 14-sentence repetition task were analyzed for prosodic and temporal features.
• Supervised machine learning models (Random Forest, KNN, SVM) were trained and validated for classification.

Main Results:

• Random Forest model achieved 91% cross-validation accuracy in differentiating the three groups.
• Speech-based classification accurately identified clinical phenotypes and underlying pathologies.
• The study demonstrated speech markers as valuable tools for early prognostic identification.

Conclusions:

• Speech-based classification offers early prognostic insights by linking clinical phenotypes with CSF profiles.
• Enhanced prognostic precision through speech analysis can improve early intervention strategies.
• Larger studies are needed to validate speech classification as a reliable non-invasive tool for differential prognosis.