The trajectory of cognitive decline, once established, is not fixed. It bends in response to interventions. But it does not bend equally well at every point along the lifespan. The acceleration window — the period during which prevention work has the highest yield per unit of intervention — is roughly age 40 to 70, with the most leverage in the 50s and 60s. By the time clinical dementia has been diagnosed, the trajectory has narrowed. Interventions that would have prevented or delayed the disease now serve mostly to slow its progression rather than redirect it.
The Lancet Commission’s 40% population-attributable risk figure — the claim that approximately 40% of dementia cases are theoretically preventable through modification of the twelve risk factors — applies most cleanly to this window. The figure is an aggregate that assumes the modifications happen while the trajectory is still bendable. A 75-year-old who quits smoking has gained brain-protective benefit; a 55-year-old who quits has gained substantially more. The same logic applies to every risk factor: earlier modification, larger effect.
The acceleration window is not equally open for everyone. Patients with APOE ε4 alleles, family history of early-onset dementia, persistent psychiatric illness, or accumulated risk factors enter the window earlier and may have a narrower opening — for them, action at 45 may be more important than action at 65. Patients with strong cognitive reserve from education, complex work, multilingualism, and lifelong learning have wider openings and more buffer. The window is individually calibrated, and the calibration is part of the longevity-psychiatry assessment.
The clinical implication is that the comprehensive longevity-psychiatry workup is most appropriately delivered to the 45-65 age range, when the most can still be changed. Earlier work is genuinely preventive — building cognitive reserve, establishing baselines, treating risk factors before they accumulate. Later work shifts toward treatment of established disease. Both matter; the leverage is different, and the work in the acceleration window is the work that produces the population-level effect.
What "bending the trajectory" actually looks like clinically: hearing loss treated at 55 instead of 75 saves a decade of additional cognitive risk exposure during the very years when amyloid is accumulating. Sleep apnea treated at 50 prevents chronic hypoxic insult through the decade when glymphatic clearance most matters. Depression treated successfully and durably at 45 prevents the accelerated brain aging — telomere shortening, hippocampal volume loss, sustained inflammation — that untreated depression produces. Hypertension and apoB controlled in the 50s reduces small vessel disease that would otherwise contribute to vascular cognitive impairment in the 70s. Each is a real intervention with real evidence. None dramatic on its own. Cumulatively, when applied across the window, they produce the bend.
The work is not glamorous. It is the disciplined application of evidence-based interventions over the window when they still work, in patients who are largely asymptomatic and who therefore feel no urgency. This is the structural problem: the patients in the acceleration window do not feel that they are in any window at all. They feel fine. The clinical task is to deliver care that they do not yet feel they need, for an endpoint they cannot yet see, on a timeline measured in decades.
The window closes. The 80-year-old with established mild dementia is no longer in the acceleration window. The work shifts from prevention to symptom management, family support, and dignified care. The work that mattered most for this patient happened (or did not happen) twenty years earlier — when it was unglamorous, when the patient was asymptomatic, and when the medical system around them was watching different things.