Progressive supranuclear palsy, a relentless brain disease first described in the 1960s by neurologists Steele, Richardson, and Olszewski, has long been considered a rare disorder, with prevalence estimates typically hovering around one in ten thousand people. A sweeping new analysis of autopsy data from 38 U.S. Alzheimer’s Disease Research Centers, published in Acta Neuropathologica, now suggests that this reputation for rarity may be profoundly wrong. The study, led by Benjamin J. Cushing, Ryan K. Shahidehpour, and Peter T. Nelson of the University of Kentucky together with Gabor G. Kovacs of the University of Toronto, examined 6,845 brain donors and identified 297 cases with autopsy-verified PSP pathology. What they found challenges both the epidemiology and the clinical picture of a disease that may be silently affecting far more aging brains than anyone realized.
PSP belongs to a family of neurodegenerative conditions called tauopathies, in which a protein called tau, normally a stabilizer of microscopic transport tubes inside neurons, misfolds and clumps into toxic aggregates. In PSP, the misfolded tau is of the four-repeat variety and accumulates in characteristic lesions: neurofibrillary tangles inside neurons, tufted astrocytes in supporting glial cells, and coiled bodies within oligodendrocytes. Classic PSP, known as Richardson syndrome, announces itself with paralysis of vertical eye movements, pseudobulbar palsy, slurred speech, and dystonia. Yet the new data confirm what specialists have increasingly suspected: pure Richardson syndrome accounts for less than a quarter of clinical PSP, and the disease wears many disguises, including a parkinsonism-predominant form and at least seven other described subtypes.
The research team capitalized on the National Alzheimer’s Coordinating Center Neuropathology Data Set, which aggregates longitudinal clinical assessments and autopsy findings from federally funded research centers across the United States. Participants were evaluated roughly annually using standardized Uniform Data Set forms capturing cognition, medications, neurological signs, and movement symptoms. Crucially, the investigators stratified cases by cognitive status at recruitment: people who entered the studies with dementia or clinical PSP were labeled prevalent cases, while those who joined while cognitively normal and later developed disease were incident cases. Because definitive PSP diagnosis requires autopsy, and autopsy rates have been declining for decades, this longitudinal design offered a rare window onto the true burden of the disease in a research population.
The headline numbers are striking. Among participants recruited while cognitively normal, 3.2 percent had PSP pathology at autopsy, a frequency the authors describe as higher than expected from prior epidemiological studies. Among those recruited with dementia or clinical PSP, 4.5 percent carried the diagnosis. Both figures dwarf the prevalence of other frontotemporal lobar degeneration tauopathies in the same cohort: corticobasal degeneration was found in 0.7 percent of initially normal donors and Pick’s disease in a mere 0.1 percent. In other words, PSP was the most common FTLD-tau subtype in this large sample, appearing in roughly one in thirty older research volunteers who donated their brains, a proportion that sits uncomfortably beside the one-in-ten-thousand figure that has anchored textbooks for years.
Just as revealing is how differently incident and prevalent PSP behaved. The 47 incident cases, people who entered research cognitively normal and later developed PSP, died at an average age of 89.5 years, more than thirteen years older than the 76.3-year average among the 193 prevalent cases. Only 6.4 percent of incident cases had been clinically diagnosed with PSP before death, compared with 48.7 percent of prevalent cases. Instead, the incident group was more than three times as likely to have received a Parkinson disease diagnosis, and they were significantly less likely to show the stereotypical PSP features of eye movement changes, dysarthria, axial rigidity, gait disorder, and apraxia of speech. More than half of the incident cases progressed to mild cognitive impairment or dementia before dying, suggesting that in older adults PSP often masquerades as something else entirely.
To probe the biology underlying these clinical differences, the team turned to artificial intelligence-assisted digital pathology on a subsample of 23 PSP cases from the University of Kentucky. Brain sections from the dorsolateral prefrontal cortex, the putamen, and the midbrain were immunostained for hyperphosphorylated tau using the PHF-1 antibody, digitized at 40-times magnification, and analyzed with the HALO image analysis platform. The researchers trained machine-learning classifiers on hundreds of thousands of iterations to segment tissue, exclude artifacts, and then detect and phenotype every tau-positive structure as a tangle, a tufted astrocyte, or a coiled body. The phenotyping classifier achieved an F1 score of 0.983, with class-specific scores above 0.98, indicating near-perfect agreement with expert manual annotation and lending considerable confidence to the quantitative results.
The digital pathology findings revealed a coherent anatomical story. Across regions, prevalent PSP cases generally carried heavier tau burdens than incident cases, with the most pronounced separation in the frontal cortex, where prevalent disease showed elevated densities of both coiled bodies and neurofibrillary tangles. In the putamen, the composition of pathology differed in a telling way: incident cases had a proportionally higher contribution of tufted astrocytes, while prevalent cases were dominated by neuronal tangles, differences that reached statistical significance. The midbrain, one of the earliest and most consistently affected regions in PSP, showed substantial tau pathology in both groups without significant differences, consistent with the idea that all PSP cases share a common core of brainstem and basal ganglia involvement, and that the groups diverge mainly in how far the disease has spread beyond those epicenters.
These patterns fit neatly within a staging framework proposed by Kovacs and colleagues, in which neuronal tau first accumulates in the pallido-nigro-luysian axis, striatal astrocytes follow, and striatal tangles plus widespread cortical involvement emerge later. Incident cases, with their astrocyte-heavy striatal pathology and lighter cortical burden, may simply sit at an earlier stage. But the thirteen-year gap in age at death complicates a purely stage-based interpretation. The authors raise the possibility that at least some incident PSP represents a slowly progressive, protracted form of the disease, in which misfolded tau behaves in a more benign fashion, seeding pathology that accumulates over decades without producing classic symptoms. Recent work on four-repeat tau seeding activity suggests molecular subtypes of PSP may exist, and the observation that group differences were driven by neuronal rather than astrocytic tau echoes findings in corticobasal degeneration, where neuronal tau predominance signals more aggressive disease.
The study also flagged intriguing epidemiological signals. People with PSP pathology were more likely to have baseline depression than those without it, 45.1 percent versus 35.0 percent, and less likely to report a first-degree relative with cognitive impairment, 38.0 percent versus 49.7 percent. Participants with incident PSP were more likely to have comorbid Lewy body pathology at autopsy, whereas PSP cases overall were less likely than non-PSP cases to carry Lewy bodies. These associations hint at genetic and environmental influences on PSP risk that remain poorly understood, and the authors are careful to note that their cohort, drawn from Alzheimer’s-oriented centers that skew toward White, female, highly educated participants, is not population-representative. Movement disorder specialists may not have been involved in every diagnostic evaluation, and some subtle parkinsonian signs could have been missed or documented inconsistently.
Even with those caveats, the implications are hard to ignore. If PSP pathology is present in roughly three percent of cognitively normal older adults who come to autopsy, and if most of those cases were never recognized in life, then a substantial reservoir of undiagnosed tauopathy exists in the aging population, with consequences for clinical trial design, biomarker development, and the counseling of families. The authors emphasize that they are not proposing a new classification system, and that the concept of incident PSP would manifest differently in other cohorts. But their message is unambiguous: PSP may have a larger public health footprint in the elderly than commonly appreciated, its presentation in advanced age is often atypical, and pairing longitudinal clinical cohorts with AI-driven quantitative neuropathology offers a powerful new lens for catching the disease before it hides in plain sight.
Subject of Research: Epidemiology and neuropathology of progressive supranuclear palsy, a tau protein neurodegenerative disease of aging
Article Title: Incident versus prevalent progressive supranuclear palsy: analyses of U.S. Alzheimer’s Disease Research Centers data indicate clinical and neuropathological features of a relatively common and underrecognized tauopathy in the aging brain
Article References: Cushing, B. J., Shahidehpour, R. K., Abner, E. L., Jicha, G. A., Neltner, J. M., Breig, M. K., Teodorescu, K. L., Neltner, A. M., Coskun, E. P., Kovacs, G. G., & Nelson, P. T. (2026). Incident versus prevalent progressive supranuclear palsy: analyses of U.S. Alzheimer’s Disease Research Centers data indicate clinical and neuropathological features of a relatively common and underrecognized tauopathy in the aging brain. Acta Neuropathologica, 152(1), Article 51. https://doi.org/10.1007/s00401-026-03100-8
Image Credits: AI Generated
DOI: 10.1007/s00401-026-03100-8
Keywords: progressive supranuclear palsy, tauopathy, neurodegeneration, autopsy, National Alzheimer's Coordinating Center, digital pathology, artificial intelligence, frontotemporal lobar degeneration, parkinsonism, neurofibrillary tangles, tufted astrocytes, aging brain
News Source: Diana Fleming. (October 13, 2026). Hidden Tauopathy: Autopsy Data Reveal Progressive Supranuclear Palsy Is Far More Common Than Thought. Scienmag.
