The distribution of neurofibrillary tangles (NFTs) and neuritic plaques (NPs) was mapped in 39 cortical areas of 11 brains of patients with Alzheimer's disease (AD). Whole hemisphere blocks were embedded in polyethylene glycol (Carbowax), sectioned coronally, and stained with thioflavin S and thionin. The densities of NFTs and NPs were assessed using a numerical rating scale for each area. Scores were grouped by type of cortex and by lobe for statistical analysis. Highly significant differences were ohtained. For example, limbic periallocor tex and allocortex had more NFTs than any other type of cortex. In descending order, the density of NFTs was as follows: penallocortex (area 28) > allocortex (suhiculum/CA1 zones of hippocampal formation, area 51) > corticoid areas (accessory basal nucleus of amygdala, nucleus hasalis of Meynert) > proisocortex (areas 11, 12, 24, 23, anterior insula, 38, 35) > nonprimary association cortex (32, 46, superior temporal sulcus, 40, 39, posterior parahippocampal cortex, 37, 36) > pri mary sensory association cortex (7, 18, 19, 22, 21, 20) > agranular cortex (44–5, 8, 6, 4) > primary sensory cortex (41–2, 3–1–2, 17). The laminar distribution of NFTs tended to be selective, involving primarily layers III and V of association areas and layers II and IV of limbic periallocortex. There were far more NFTs in both limbic and temporal lobes than in frontal, panetal, and occipital lobes. In general, NPs were more evenly dis tributed throughout the cortex, with the exceptions of limbic periallocortex and allocortex, which had notably fewer NPs than other cortical areas. Temporal and oc cipital lobes had the highest NP densities, limbic and frontal lobes had the lowest, and panetal lobe was intermediate. No significant left-right hemispheric dif ferences for NFT or NP densities were found across the population, and there was no relationship between du ration of illness and densities of NFTs or NPs.

The regional and laminar distribution of NFTs (and, to a lesser degree, that of NPs) suggests a consistent pattern of vulnerability within the cerebral cortices that seems correlated to the hierarchies of cortico-cortical connections. The higher-order association cortices, es pecially those in the anterior and ventromedial sectors of temporal lobe, are the most vulnerable, while other cortices appear less vulnerable to a degree commen surate with their connectional “distance” (i.e., synapses removed) from the limbic areas.