PROJECT SUMMARY Primary progressive aphasia (PPA) is a clinical syndrome characterized by isolated, progressive loss of speech and language abilities due to frontotemporal lobar degeneration (FTLD) or Alzheimer’s disease (AD). Three clinical variants of PPA have been identified: i) logopenic variant (lvPPA) associated with loss of phonological abilities, left temporal-parietal atrophy and atypical AD pathology; ii) nonfluent variant (nfvPPA) with motor speech/grammar deficits, left inferior frontal damage and FTLD pathology; and iii) semantic variant (svPPA), with conceptual knowledge loss, anterior temporal damage and mostly FTLD-type pathology. In the first funding period, we demonstrated distinct functional neuropathophysiology across PPA variants (>40 pubs). Building upon these findings, in this renewal, we focus on revealing important mechanistic neural circuit abnormalities in PPA variants. Such mechanistic understanding is the foundation for development of better neuromodulatory or behavioral interventions for PPA. We leverage the unmatched temporal resolution of magnetoencephalography imaging (MEGI) with structural and diffusion MRI (to account for neurodegeneration of grey and white matter), spectral graph modeling, speech motor control modeling, machine learning, as well as detailed cognitive and language phenotyping. The specific aims are: 1) To determine distinct resting-state structure-function imaging of neural oscillations in PPA; 2) To examine mechanisms of control and learning in the speech production in PPA; 3) To examine neural interactions between speech production and semantic representations in PPA. This unique combination of multimodal brain imaging (MEGI, structural MRI & diffusion MRI), modeling (spectral graph, speech motor control), machine learning, and speech neuroscience will: a) delineate the functional manifestations of brain network dysfunction in PPA variants, and b) identify putative brain network targets and strategies for behavioral, or neuromodulation therapies for PPA.