Braak Staging and NIA-AA Neuropathologic Criteria

A pathology report lists a Braak stage, a Thal phase and a CERAD score, and the reader needs one overall level. Braak staging grades tau spread by region, Thal phases grade amyloid spread, and the NIA-AA scheme combines them.

Quick answer

• Tau: the Braak stage runs from I to VI, from entorhinal and transentorhinal cortex to primary visual cortex, as summarised in this staging paper.
• Amyloid: the Thal phase runs from 1 to 5, from temporal neocortex to cerebellum.
• Plaques: the CERAD score reflects neuritic plaque density on silver-stained sections.
• Combination: each result is converted to an A, B or C score, and the three are read together as one Alzheimer’s disease neuropathologic change (ADNC) level.

This guide stays with staging and scoring. For the wider disease context, see the types of neurodegenerative diseases covered on this site.

Reading the Braak stage on a tau report

Braak staging records how far neurofibrillary tangle pathology, which is tau pathology inside neurons, has spread across the brain. The stages follow a regional order. As described in the literature on Braak tangle stage and Thal amyloid phase, the earliest sites of tau pathology lie in the entorhinal and transentorhinal cortex (stage I). Tau then spreads to the hippocampus (stage II) and the temporal cortex (stage III). It eventually reaches other regions of cerebral cortex (stage IV), and finally the visual association cortex (stage V) and the primary visual cortex (stage VI).

Because the stage is defined by anatomy, the first job when reading a report is to find which regions carry tau. The highest region reached sets the stage, so the report’s regional description matters more than a general phrase such as “widespread”.

Early stages: the transentorhinal landmark

The transentorhinal cortex is the anchor of the scale. Tau confined to the entorhinal and transentorhinal cortex corresponds to stage I. Once the hippocampus is also involved, the case has moved to stage II. A report that names only these medial regions describes the lowest part of the scale, however clearly the tangles stand out under the microscope.

Later stages: from temporal cortex to visual cortex

Each higher stage names a wider cortical territory. Stage IV is deliberately broad, “other regions of cerebral cortex”, so it is the stage where a report’s regional detail is most useful to read. The top of the scale is defined by the visual cortex, and the distinction between V and VI is the distinction between association cortex and primary cortex. When deciding between the last two stages, check which part of the visual cortex carries tau.

Reading the Thal phase on an amyloid report

The Thal phase grades the spread of amyloid deposition, and it follows a different route from tau. In the same staging description, amyloid commences within neocortical regions, particularly the temporal lobe (phase 1). It then spreads to allocortical regions such as the hippocampus and amygdala (phase 2). Next it enters subcortical regions (phase 3), then the brain stem (phase 4), and finally the cerebellum (phase 5).

The contrast with tau is practical. Tau starts in the entorhinal and transentorhinal cortex, while amyloid begins in neocortex, so a low Thal phase and a low Braak stage describe different anatomy. In staging tables the labels are compact: Thal means the Thal phase of amyloid deposition, and Braak means the Braak tau stage. Check which label a number belongs to before combining anything.

Braak stages and Thal phases side by side

This table puts both scales in one place so a regional finding can be matched to its step.

Regions and ordering in the table come from the Braak tangle stage and Thal amyloid phase paper.

Combining the scores into an ADNC level

The NIA-AA approach does not rely on one measurement. It reads amyloid spread, tau spread and neuritic plaque density together and reports a single ADNC level. The sequence below keeps the inputs separate until the last step.

1. Grade amyloid spread. Assign the Thal phase from the regions that carry amyloid.
2. Grade tau spread. Assign the Braak stage from the regions that carry neurofibrillary tangles.
3. Grade plaque density. Assign the CERAD score from neuritic plaque density on silver-stained sections.
4. Convert to A, B and C. The amyloid result gives the A score, the tau result gives the B score and the plaque result gives the C score.
5. Read the combination. Match the A, B and C scores against the NIA-AA combination table to obtain the ADNC level, rather than relying on a rule of thumb.

A short illustration shows why the inputs must stay distinct. Suppose amyloid is found only in temporal neocortex and tau only in the entorhinal and transentorhinal cortex. The amyloid result is then Thal phase 1 and the tau result is Braak stage I, and both still need their A and B conversion before the level is read. A case with tau in primary visual cortex and amyloid in the cerebellum would be graded the same way, from the highest region reached on each scale. The cutoffs that turn each phase, stage and plaque grade into a score, and the final lookup, are those printed in the NIA-AA guideline.

The lesions behind these scores are described in our alzheimer’s disease neuropathology guide. The practical lesson here is to record all three results, not just the highest stage, because the level depends on their combination.

Braak stage and clinical dementia

A Braak stage describes where pathology has spread, not whether a person had dementia. The clearest supported evidence on this point comes from Parkinson’s disease, which has its own Braak scheme for synuclein pathology. That scheme is a separate series of discrete stages and should not be confused with the tau stages above. Our guide to the neuropathology of parkinson’s disease covers that disease in detail.

• Cognition tracks the later stages in group data. In Parkinson’s disease, average MMSE scores consistently decreased from Braak stages 3 to 6. In the staging description, stage 5 includes the onset of cognitive impairment and by stage 6 decline may have progressed to Parkinson’s disease dementia.
• Motor severity does not follow the stage. A critical evaluation in Annals of Neurology (2008) found no relation between Braak stage and the clinical severity of Parkinson’s disease.
• A high stage does not guarantee symptoms. The same review notes that aged people without neurological signs can have brain synucleinopathy ranging up to Braak stages 4 to 6 at postmortem.

The decision for readers is to treat a stage as a statement about distribution. It supports a clinical conclusion only when it is read alongside the clinical history.

When a case does not follow the expected order

Staging schemes summarise common patterns, and some cases depart from them. Check these situations before forcing a case into a stage.

Tau outside the expected starting point

The staging order has been challenged by tau found in brainstem nuclei in young people. A study of 42 cognitively normal individuals under 30 years of age reported this finding. When a report describes tau in unexpected places, describe the regions as found and then assign the stage on the scale’s own regional definitions.

Synuclein patterns that do not fit

The Parkinson’s scheme has similar limits. Less than half of post-mortem Parkinson’s disease cases showed a close fit to it. The Annals of Neurology review adds that the patterns of synucleinopathy Braak described are often not observed, particularly in dementia with Lewy bodies. If a case does not fit, state that it does not fit instead of assigning the nearest stage.

Frequently asked questions