Primary Age-Related Tauopathy Explained

These points come from a University of Texas Medical Branch research record of a peer-reviewed study of PART tissue.

Primary age-related tauopathy (PART) is a pattern of tau pathology restricted to the hippocampus in the absence of amyloid-β. The label depends as much on what is missing as on the tangles that are present.

PART is one of the types of neurodegenerative diseases that neuropathologists separate by protein signature. It is easiest to understand through what a diagnosis changes, so the consequences come first and the mechanism after.

What a PART finding means in practice

A PART finding tells you which protein is abnormal and which is not. Tau has aggregated, mainly in the hippocampus, and amyloid-β is absent. That combination is the whole definition, so the amyloid result carries as much weight as the tau result.

For a family reading an autopsy report, or a student reviewing a case, the practical question is whether the report documents both halves. A statement of hippocampal tau alone is incomplete. A statement of hippocampal tau with a negative amyloid-β stain is what the PART definition describes.

Tau pathology is usually described in terms of neurofibrillary tangles. A neurofibrillary tangle is an intraneuronal aggregate of tau protein. In PART these tangles are the defining lesion, not a companion to plaques.

Why hippocampal restriction matters

The hippocampus is a medial temporal structure involved in forming memories, and it is where the PART definition places the tau pathology. Restriction to this region gives the diagnosis its boundaries. If tau had spread widely across the cortex, the case would no longer match the narrow pattern the term describes.

Braak staging is the usual shorthand for how far tangles have spread through the brain. Because PART is defined as hippocampal-restricted, a PART Braak stage sits at the early, medial temporal end of that scale. The restriction is part of the definition, not a stage that PART eventually outgrows.

PART versus early Alzheimer disease

The deciding difference is amyloid-β. In PART the hippocampal tau pathology occurs without amyloid-β, whereas Alzheimer disease is characterised by amyloid-β together with tau. The detailed lesion profile is laid out in our guide to alzheimer’s disease neuropathology.

This makes the two easy to separate on paper and harder to separate in thought. Early Alzheimer disease also involves medial temporal tau. A pathologist who sees only hippocampal tangles must therefore check the amyloid-β immunostain before choosing a label. The stain is the step that moves a case from one category to the other.

• Amyloid-β negative, hippocampal tau: matches PART as defined.
• Amyloid-β present with tau: belongs with Alzheimer-type pathology, not PART.
• Incomplete staining or sampling: leaves the category uncertain, so a confident PART label is not justified.

Tau pathology also appears in other conditions. If you want a different tauopathy to compare against, chronic traumatic encephalopathy pathology is a useful contrast, because the label there rests on a different clinical and pathological context.

Does PART cause cognitive decline?

The evidence supports an association with altered synaptic function, not a settled causal answer. In a study of PART tissue, hippocampal accumulation of oligomeric and paired helical filament (PHF) tau was negatively correlated with the synaptic excitation-to-inhibition ratio, abbreviated sE/I. Oligomeric and fibrillar tau were associated with a shift toward synaptic inhibition, predominantly within the hippocampus. The work is summarised in the university research record.

A shift toward inhibition in a memory-related region is a plausible route to cognitive effects. But a correlation in post-mortem tissue does not show that a given person had symptoms or that tau caused them. The sensible reading is that PART should not be dismissed as functionally silent, and that its clinical weight is something to establish case by case.

Where the tangle-predominant label fits

Tangle-predominant dementia is the term used when tangle pathology dominates and dementia is present. It shows why the cognition question matters for tangle-only cases. If tangles alone can be associated with cognitive change, the amyloid-negative pattern cannot be treated as automatically harmless. How the term relates to PART is part of the wider classification debate below.

Separate disease or early Alzheimer-spectrum stage?

Defining PART by the absence of amyloid-β invites a question that the definition itself cannot answer. Is a tangle-only brain a distinct disease, or a point on a path that may later acquire amyloid-β? The evidence available here describes the PART pattern and its synaptic correlates. It does not show how a given case would evolve over time.

That leaves two reasonable working positions, and the choice depends on your purpose.

• Treat PART as its own category when you are classifying a finished autopsy. The definition is explicit, and the label records exactly what was and was not seen.
• Keep the Alzheimer-spectrum question open when you are reasoning about progression. A single post-mortem snapshot cannot show where a brain was heading.

The two positions do not conflict. One concerns what a specimen shows, and the other concerns what that pattern might mean over a lifetime.

How PART has been studied

PART research relies on autopsy tissue. A study published on Feb 9, 2026 in the Journal of Alzheimer’s disease used autopsy-derived hippocampal and superior middle temporal gyrus tissue from neuropathologically validated PART specimens. Validation matters because the category is defined by tissue findings, so a misclassified specimen would blur the very contrast the study set out to measure. The details are on the publication record.

Comparing the hippocampus with the superior middle temporal gyrus lets researchers ask whether an effect is local. The study reported that the tau and synaptic associations were predominantly hippocampal, which fits the regional restriction in the definition.

Measuring tau and synaptic function

The researchers quantified tau species by western blot, including monomers, oligomers and paired helical filaments. Separating these forms matters because tangles are only the most visible end of tau aggregation. Smaller aggregation-prone assemblies were the ones linked to synaptic changes.

Synaptic function was assessed by microtransplantation of synaptosomal membranes into Xenopus laevis oocytes. This approach lets human synaptic membranes from post-mortem tissue be tested for their excitatory and inhibitory behaviour. It is an indirect, laboratory-based readout, so it supports an association with tissue function without showing what happened in a living patient.

Tau assemblies and synaptic balance

The study also examined which proteins track with reduced sE/I. Proteins linked to it were enriched for pathways related to vesicle-mediated transport, synaptic endocytosis and neurotransmitter receptor regulation. These are processes by which synapses package, recycle and respond to signalling molecules.

The implication is that tau assemblies in PART may be tied to the machinery of the synapse, not only to the visible tangle. That is a hypothesis-generating finding. It suggests where to look next, and it does not yet define a treatment target.

Using this information well

How you apply PART knowledge depends on your role.

• For students and trainees: learn the definition as a pair of conditions, hippocampal tau present and amyloid-β absent, and practise checking both before naming the pattern.
• For families reading a report: ask whether the amyloid-β stain was reported and what the tau distribution was. A PART description without an amyloid result is not complete.
• For readers following the research: weigh synaptic findings as associations from validated autopsy tissue, and wait for longitudinal evidence before drawing conclusions about progression.
• For anyone comparing diagnoses: use PART as a contrast case. It shows what tau pathology looks like when amyloid-β is not part of the picture.

Frequently asked questions