Researchers measured the intrinsic circadian period of skin fibroblasts cultured from cognitively impaired older adults enrolled in the BICWALZS cohort, and found that a longer cellular clock cycle was associated with higher plasma pTau-217, neurofilament light (NfL) and GFAP, along with greater medial temporal lobe atrophy on brain imaging, according to a paper published in the Proceedings of the National Academy of Sciences on March 2, 2026. The work, led by Hyun Woong Roh and colleagues, isolated skin cells from study participants and tracked their molecular clock cycling in the lab over multiple days, producing a per-person period length that could then be compared against blood biomarkers and brain scans collected from the same individuals. The approach differs from prior circadian research in dementia, which has mostly relied on reported symptoms such as sundowning or wearable-tracked sleep fragmentation; this study instead measures a cell-autonomous property that persists once tissue is removed from the body and grown in a dish, and the authors report it tracked with markers tied to tau pathology, neurodegeneration and glial activation rather than with amyloid alone. The findings come from a single observational cohort of older adults already showing cognitive impairment, so the paper documents an association among people who already have symptoms, not a test that predicts who will develop them, and the authors did not report on cognitively unimpaired comparison participants in this analysis. A companion PNAS commentary published April 6, 2026, framed the work as raising cellular circadian timing as a candidate marker worth further study, while noting it has not been validated as a diagnostic or prognostic tool.