This hypothesis has stuck with me: early-onset tauopathy in the entorhinal cortex might not be pathology at all. It could be a programmed diversification strategy.

We operate on an assumption so deeply embedded we don't question it: anyone developing cognitive decline before age 75 is failing a universal biological standard. But evolutionary biology doesn't deal in universals. Differential mortality isn't noise — it's the raw material selection actually works on.

Consider what the entorhinal cortex does. It's the brain's GPS. Grid cells, border cells, head direction cells — this is the navigation system that gets you home drunk, lets you visualize your kitchen from memory, and helps you plan routes. It's also where tau pathology first appears, often decades before any clinical symptoms show up.

What if early grid cell degradation isn't cellular failure? What if it's a controlled shutdown — a biological signal saying "this individual will now contribute differently to the population"? Not as a navigator, but perhaps as a mentor, a storyteller, a keeper of social knowledge that doesn't require spatial precision?

The data already shows this: not all cognitive decline is equal. Some people lose navigational ability but keep verbal fluency. Some lose executive function but retain emotional regulation. We treat all of it as loss. But what if it's redistribution?

I want to test this directly. Map tau propagation patterns in early-onset cases versus typical onset. Compare which functional domains degrade first. Ask whether there's a predictable sequence that differs by decades — and whether that sequence correlates with different life-history strategies.

This isn't about giving up on treatment. It's about asking whether we've been fighting a shadow. If early tauopathy is partially canalized as a biological strategy, we need to know that before spending another billion trying to reverse something that might be self-limiting.

Looking for: collaborators with access to longitudinal tau-PET datasets, evolutionary biologists willing to model differential senescence, and anyone funding who wants to question the assumption that longer always means better.

Who's in?