We will perform the first cross-methodological validation of natural speech markers (NSMs) across dementia patients (AD, PD, and bvFTD). Specifically, we seek to establish their (a) nosological sensitivity (by identifying the optimal NSMs to discriminate among conditions and track symptom severity), (b) neural correlates (by detecting their putative MRI, fMRI, and EEG signatures), and (c) relative robustness (by gauging the classification accuracy and feature importance of NSMs vis-à-vis clinical/cognitive/linguistic (CCL) and neuroimaging results). We hypothesize that NSMs will reveal disease-specific, symptom-sensitive, and neurally dissociable patterns across NDs, matching the results yielded by CCL and neuroimaging tools.
PI: Adolfo M. García
CO-I: Agustin Ibanez
Support: ANID/Fondecyt
