Large-scale EEG network alterations in late-onset epilepsy of unknown etiology: A comparative study with Alzheimer's disease

Late-onset epilepsy of unknown etiology (LOEU) is frequently accompanied by cognitive impairment and has been increasingly linked to Alzheimer's disease (AD), prompting questions about overlapping versus distinct pathophysiological mechanisms. While both conditions have been associated with cerebral small vessel disease (CSVD) and large-scale functional network dysfunction, the shared and disorder-specific network substrates underlying cognitive decline remain incompletely characterized. In this prospective, two-center cohort study, we used biomarker-supported AD as a clinical reference point to investigate network alterations in LOEU. We systematically assessed CSVD burden and performed resting-state electroencephalography (EEG)-based functional connectivity analysis in 60 patients with LOEU, 65 patients with AD, and 60 cognitively healthy controls. Functional connectivity was quantified using the weighted phase lag index (wPLI). Compared with healthy controls, both patient groups exhibited higher CSVD burden and shared significant alpha-band desynchronization, whereas significant theta-band hypersynchronization and beta-band desynchronization were exclusive to AD. Furthermore, multivariate analyses revealed distinct substrates underlying cognitive decline. In LOEU, cognitive impairment was independently associated with both vascular burden (specifically white matter hyperintensities) and global alpha-band desynchronization. In contrast, cognitive decline in AD was exclusively associated with specific oscillatory abnormalities (concurrent theta-band hypersynchronization and alpha-band desynchronization). These findings indicate that while alpha-band disruption serves as a shared downstream network manifestation, cognitive decline in LOEU possesses a unique, vascular-associated pathophysiological profile distinct from Alzheimer's-type neurodegeneration. Overall, these results highlight the value of integrating EEG and structural markers to decouple shared and disorder-specific mechanisms, providing functional insights to further inform the clinical evaluation of LOEU.