Timely given the current pandemic, we present epidemic spread of tau through the human brain. Instead of a virus traveling between cities using highways, we model tau traveling through brain regions using neuronal connections, in Alzheimer's disease https://rdcu.be/b4pQd  1/

A great collaboration between @biofinder_study and @MCINation , featuring @MedinYasser , @LevitisLiza @acehigh1952 and others! Some highlights in this thread: 2/

The epidemic spreading model was developed by @MedinYasser and described here: https://tinyurl.com/y9mesnhf . We employ this method to model the diffusion and propagation of tau-PET through the cerebral cortex across DWI- and rsfmri-measured brain connections. 3/

The simulated spread of tau from the entorhinal cortex through cortical connections matches well the observed distribution of tau-PET in the brain. The entorhinal cortex proved to be the best model epicenter, matching expectations from autopsy studies 4/

Surprisingly, we found the model performed well in Aβ-negative individuals with very low levels of tau, perhaps picking up the ubiquitous age-related phenomenon known as primary age-related tauopathy (PART). 5/

However, Aβ seems to accelerate tau spread. Regions with more Aβ showed more tau-PET signal than would be expected given their anatomical connection to the entorhinal cortex. 6/

For all you tau-PET folks out there, we find that using mixture-modeling to identify “positive” tau-PET signal improves model performance, and produces a more “Braak-like” distribution of tau across the brain. 7/

We also examine in detail how different modeling and preprocessing choices influence our results. Choice of regions and data representation have a lot of influence, preprocessing choices and connectivity have less, and PVC and cohort selection have hardly any. 8/

Excellent previous papers by e.g. @nfranzme , @RikOssenkoppele and others show evidence for neuronal spread of tau in humans. We expand on this evidence through experimental simulation, concluding tau spreads through the MTL-network in aging before being accelerated by Aβ in AD.