According to Forbes, AI is fundamentally changing how we understand and treat Alzheimer’s disease, which affects over 7 million Americans and strikes about 1 in 9 people aged 65 and older. MIT professor Manolis Kellis explains that AI has shifted research from correlation models to causation models, analyzing thousands of post-mortem brain samples. The breakthrough finding shows Alzheimer’s involves decreased myelination around neurons and disrupted lipid transport, where lipids get stuck in cellular structures causing damage. Researchers identified specific disrupted pathways in myelination, cholesterol transport, and neuroinflammation that recur across different patients. The approach uses generative AI to design new therapeutics tested on human brain organoids created from patients’ own skin cells. This could lead to highly personalized Alzheimer’s treatments targeting individual biological pathways.
What we’re actually learning about Alzheimer’s
Here’s the thing about Alzheimer’s – we’ve been looking at it wrong for decades. The amyloid plaque theory has dominated, but this new research suggests the real action is happening with the myelin sheath that protects neurons. Basically, the oligodendrocytes that maintain this protective coating stop transporting lipids properly. Those lipids then accumulate and cause damage from the inside.
What’s fascinating is they found people with plenty of amyloid buildup but no cognitive decline at all. That tells you there are protective pathways we haven’t understood until now. The AI models are identifying exactly which biological systems are breaking down – and it’s a surprisingly small number of recurring pathways across different patients.
How AI is revolutionizing drug discovery
This isn’t just about understanding the disease better – it’s about creating treatments in ways that were literally impossible before. Kellis describes embedding every chemical ever synthesized into a “chemical landscape” where AI can explore neighborhoods of chemical function. They’re basically using generative AI to design drugs that target specific disrupted pathways.
And here’s where it gets really sci-fi: they’re testing these drugs on brain organoids grown from patients’ own skin cells. We’re talking about creating 100 personalized “brains” from 100 different patients and testing 100 drugs on each. They’re measuring everything from gene expression to how neurons fire. It’s personalized medicine at a scale we’ve never seen.
What this means for prevention and treatment
So what about the egg connection? It turns out choline metabolism is crucial here. Eggs are the top food source of choline, which is a precursor to acetylcholine and plays a role in cell membranes. Moderate choline intake appears linked to reduced dementia risk. It’s not that eggs are some magic bullet, but they’re part of the lipid transport story that’s central to this new understanding.
The real breakthrough is that we’re moving from treating symptoms to addressing root causes. If Alzheimer’s manifests differently in different people because of which specific pathways are disrupted, then one-size-fits-all treatments were never going to work. Now we can actually look at an individual’s biological profile and say “okay, YOUR Alzheimer’s is primarily about lipid transport issues, so here’s the drug that targets that.”
Where this is heading
We’re still in early days, but the implications are enormous. Being able to test drugs on human brain organoids before they ever reach clinical trials could dramatically accelerate treatment development. And the personalization aspect means we might finally move beyond the disappointing clinical trial results that have plagued Alzheimer’s research.
The most exciting part? This approach isn’t limited to Alzheimer’s. The same AI-driven, organoid-testing methodology could transform how we understand and treat countless other diseases. We’re basically watching medicine shift from being reactive to being predictive and precise. And honestly, it’s about time.
