|
|
| |
| The brain changes before the plaques arrive |
| Alzheimer’s has a long, quiet head start. Researchers at the University of Oslo scanned healthy older adults for close to twenty years. The same people, again and again. Some of them eventually showed amyloid plaques — the sticky buildup linked to Alzheimer’s. The surprise sat in the earlier scans. |
| Let me explain what they measured. Your brain has an outer layer, the cortex. A scan can measure its thickness. In the people who later turned plaque-positive, that layer was already thinning. It had been thinning for at least seven years. |
| So the plaques were not the opening act. Something moved first. Picture a roof that leaks for years before a stain appears on the ceiling. By the time you see the stain, the water has been running a long while. Amyloid may be the stain, not the leak. |
| |
Seven years is only what they could measure. Today’s scans see plaques once they pile up. The real head start could be wider. |
|
| I could be wrong about what this means. One group of people, one measure, and thinning has other causes. But Anders Fjell, who runs the Oslo lab, drew a plain conclusion. Keep building drugs that aim at something other than plaques. I believe he is right. |
| Look at your next ten years. Not your next scan. The habits that hold that outer layer are the dull ones — blood pressure, blood sugar, sleep, movement. None of them wait for a picture. I urge you to start before anything shows up. |
| Roe, Fjell et al. · University of Oslo · Nature Neuroscience · Aug 19 2026 |
|
|
|
|
| |
| Make your memory answer first |
| Your phone remembers everything for you. Numbers, routes, birthdays, the shopping list. That is convenient. It also means you rarely pull anything out of your own head. |
| It works like this. Pulling a memory out is not just proof you have it. The pulling itself makes it stronger. Learning researchers call this retrieval practice. Reading a fact again is weak. Trying to recall it is strong. |
| Give your memory the first shot. |
| The Protocol |
| When |
| Once a day — the next time you reach for your phone to look something up |
| Duration |
| About 60 seconds a try, kept up for a month |
| Rules |
| Recall first, then check. Sit with the blank for a slow count of ten. Say the answer out loud. Once a week, drive a familiar route with the map off. |
| |
| It gets easier in about two weeks. I could be wrong about how far this carries. Retrieval practice is proven for learning, not proven to stop dementia. But it costs a minute a day. I urge you to try it. |
|
|
|
|
|
|
|
| A nightly melatonin habit came with a heart warning |
| Melatonin looks harmless. It sits beside the vitamins. Researchers pulled the records of 65,414 adults with long-running insomnia. All had taken it a year or more. Each was matched to a similar person who had not. Over five years, 4.6% of the melatonin group were diagnosed with heart failure. Among the others it was 2.7%. Call it roughly 90% higher. This is record-keeping, not a trial, so it cannot prove cause. Worth a word with your doctor if you take it nightly. |
| American Heart Association Scientific Sessions 2025 · resurfaced Aug 2026 |
|
|
| The risk does not stop at ninety |
| Ninety is not a finish line. Over 800 people joined a study at ninety or older, none with dementia. Kaiser Permanente and UC Davis tested their thinking every six months. Plenty still crossed into dementia. Women ran about twice the risk of men. One protective gene version went with roughly 60% lower risk. That part is luck, not habit. |
| Colbeth, Whitmer et al. · The Lancet Healthy Longevity · July 2026 |
|
|
| Years of money strain showed up in the brain |
| Money worry leaves a long trail. University College London followed 2,759 people born in 1946. They tracked income and hardship across whole adult lives. Then they tested thinking and scanned brains near age 71. The odd rough patch left little mark. Long stretches did — weaker scores by 53, more shrinkage by 71. The link ran stronger in the men. Treat a long squeeze as a health exposure, not just a budget problem. |
| Wels & Patalay · University College London · Innovation in Aging · July 24 2026 |
|
|
|
|
| |
| Drug hunters are chasing seventeen different targets. Not one. This year’s pipeline review counts seventeen distinct things going wrong in the brain, each with drugs aimed at it. Plaques are only one of them. Three of the stranger bets are below. |
|
| Sends insulin to the brain through the nose |
| Your brain runs on sugar. It needs insulin to use it well. In Alzheimer’s that handoff works poorly. So Wake Forest University tried feeding insulin in through the nose, where nerves carry it up to the brain. They paired it with a diabetes pill, empagliflozin. Suzanne Craft led the work. Forty-seven adults with early trouble took one, both, or neither for four weeks. The insulin arms scored better on thinking tests and showed firmer wiring on scans. The pill lowered a tangle-protein marker in spinal fluid. |
| Phase 2 · 47 adults · 4 weeks · reported Oct 2025 |
|
|
| Delivers a protective gene in one infusion |
| Some people inherit two risky copies. The gene is called APOE, and it shapes how the brain handles fat and repair. Two risky copies push Alzheimer’s odds far above average. Lexeo Therapeutics is testing a one-time infusion. It carries a protective version of that same gene into the fluid around the brain. Fifteen patients have had it. The protective version turned up in their spinal fluid, more at higher doses. Tangle-protein markers fell in nine of thirteen. No brain swelling was reported. |
| Phase 1/2 · 15 patients · interim data |
|
|
| Lowered a tangle marker with a cholesterol pill |
| This one came out of heart medicine. Obicetrapib is a daily pill built to cut cholesterol. NewAmsterdam Pharma ran it in 2,530 people with heart disease for a year. A planned side study watched Alzheimer’s blood markers. The marker for the tangle protein fell, and it fell in the highest-risk carriers too. Nobody has shown this changes memory. It is a prevention idea, not a treatment yet. |
| Phase 3 substudy · 2,530 people · 52 weeks |
|
|
