Search Abstracts | Symposia | Slide Sessions | Poster Sessions
EEG Signature of Logopenic Variant Primary Progressive Aphasia: Evidence from Resting-State Brain Rhythms
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
Kyrana Tsapkini1, Panteleimon Chriskos1, Arsenios Chriskos1, Nathan Crone1, Constantine Frangakis2; 1Johns Hopkins Medicine, 2Johns Hopkins School of Public Health
Background Logopenic variant primary progressive aphasia (lvPPA) is a neurodegenerative syndrome characterized by progressive impairments in word retrieval and sentence repetition, typically associated with degeneration of left temporoparietal language networks. Early identification of lvPPA remains challenging, and there is growing interest in developing accessible biomarkers that can complement more costly imaging modalities. Electroencephalography (EEG) provides a low-cost, non-invasive method for assessing large-scale neural dysfunction and may reveal disease-specific electrophysiological signatures. Objective: To identify resting-state EEG biomarkers that differentiate individuals with lvPPA from cognitively healthy controls and identify the electrodes contributing to these differences. Methods: Approximately three-minute eyes-closed resting-state EEG recordings were obtained from 23 individuals with lvPPA and 26 age-matched healthy controls, representing one of the largest resting-state EEG cohorts reported in lvPPA to date. EEG was acquired using a 32-channel system and underwent standard preprocessing, including filtering, independent component analysis, artifact rejection, re-referencing, and segmentation into four-second epochs. Relative power was calculated for delta (0.5–4 Hz), theta (4–8 Hz), alpha (8–12 Hz), beta (12–20 Hz), gamma (20–50 Hz), and high-gamma (50–100 Hz) frequency bands. Statistical analyses were performed using a hierarchical three-step procedure. First, frequencies associated with diagnostic group were identified using odds-ratio analyses with false discovery rate (FDR) correction1. Second, hemispheric regions contributing to significant frequency differences were identified. Third, individual electrodes responsible for the observed effects were localized. This approach extends our prior work demonstrating the utility of low-density EEG for classifying PPA and MCI and enables improved localization of disease-specific electrophysiological signatures. Results: Individuals with lvPPA demonstrated significantly increased power in the delta and theta frequency bands and reduced alpha power relative to controls. Increased delta activity was localized primarily to the left hemispheric region and was driven by the left temporoparietal electrode TP9. Elevated theta activity also originated predominantly from the left hemisphere, although no single electrode accounted for the effect. Reduced alpha power was widespread, involving multiple left hemispheric electrodes (except TP9 and CP1) and additional midline regions including Pz, Oz, and Fz and right hemisphere O2. High-gamma abnormalities further differentiated lvPPA from controls and were localized to left frontotemporal electrodes F7 and FT9, as well as posterior Oz. Conclusions: Resting-state EEG reveals a distinct electrophysiological profile in lvPPA characterized by cortical slowing, reduced alpha activity, and focal high-gamma abnormalities. Importantly, the strongest effects were localized to left temporoparietal and frontotemporal regions that correspond closely to the known neuroanatomical distribution of lvPPA pathology. These findings support EEG as a scalable and clinically accessible biomarker for lvPPA and suggest that brief resting-state recordings may provide sensitive markers of language-network dysfunction in neurodegenerative disease. Future studies with larger cohorts will determine the diagnostic utility of these markers and their relationship to disease severity and progression.
Topic Areas: Disorders: Acquired,