Selective 10-Hz photic driving deficit shared between Alzheimer's and frontotemporal dementia

<jats:p>Background. Photic stimulation is a routine clinical probe of posterior alpha circuits, yet its signature in frontotemporal dementia (FTD) and subject-level comparison with Alzheimer's disease (AD) remain uncharacterized. Methods. We analyzed two OpenNeuro datasets sharing 88 participants (86 retained: 35 AD, 29 controls [CN], 22 FTD): resting eyes-closed and photic stimulation at 5, 10, 15, 20 Hz. After ICA preprocessing we computed a photic driving index (DI), resting spectral features, and weighted phase-lag connectivity (wPLI). Linear mixed-effects models tested group effects; leave-one-subject-out cross-validation with five classifiers quantified diagnostic utility. Results. DI at 10 Hz was reduced ∼50% in AD (β = −6.96, p = 0.004) and FTD (β = −6.74, p = 0.012), robust under log, winsorized, and rank-based sensitivity analyses. Resting analyses showed 1.5 Hz IAF slowing in AD and posterior theta/alpha ratio elevation up to 6-fold (AD) and 3.3-fold (FTD). wPLI showed no group differences, but equivalence was not established (TOST p ≥ 0.16). DI split-half reliability at 10 Hz was near null (ICC = 0.06). Balanced accuracy was 0.577 for resting versus 0.543 combined (McNemar p= 0.85). Conclusion. A selective alpha-band amplitude decit is shared between AD and FTD but adds no diagnostic value over resting EEG. Distinguishing reduced cortical gain from preserved synchrony requires longer recordings powered for equivalence testing.</jats:p>