Alzheimer's

The latest sourced reporting on Alzheimer's from Dementias Today: dated, and plain about what the evidence does and does not show.

The chart is the patient now, and software has started reading it

A narrative review published on October 20 inventories where generative AI already touches Alzheimer's work. The uses furthest along are the ones no family ever sees: synthetic notes, phenotype mining from visit records, and drafted summaries. None of it has been tested in a live clinic.

Told the tone would dull them, they stopped getting better

A new review gathers the scattered research on placebo and nocebo effects in memory into one account. In a separate online trial, healthy adults aged 65 to 80 who were told an inert sound would blunt their attention stopped improving at a task everyone else improved at.

Leuven will give healthy people an inflammation, then a memory test

A triple-blind trial at UZ Leuven randomises colon-release short-chain fatty acid capsules against placebo, then injects every volunteer with bacterial endotoxin and measures what happens to working memory, inhibition and task switching. It is a cause test in a field built mostly on correlations.

The Alzheimer's risk gene that clogs the brain's wire wrappers

An MIT team knocked out every protein-coding gene, one at a time, in human oligodendrocytes carrying APOE4. The trail led to Wnt signaling, an enzyme called GSK3B, and fat droplets where insulation should be. It led there in dishes, organoids and mice, and in no patient yet.

Before the Forgetting, Something Shifts in the Gut

A narrative review published in Nutrients on 29 July 2026 gathers the evidence that short-chain-fatty-acid-producing bacteria thin out during the silent, amyloid-positive stage of Alzheimer's. It also says, plainly, that nobody yet knows which end of the story came first.

The Notebook Against the Questionnaire

Nine Italian memory clinics signed up to test whether a structured daily diary of behaviour beats standard care at lightening the load on families. Three years after the trial was registered, the results have not appeared.

A little mitochondrial stress may build a tougher brain; too much breaks it

A review accepted on 1 September 2026 in Frontiers in Molecular Neuroscience argues that mitohormesis, the idea that mild mitochondrial stress makes cells more resilient while relentless stress wrecks them, is a useful framework for linking metabolic disease to dementia risk. It is a framework, the authors stress, not a proven mechanism.

Pilot Trial: App-Guided Care Plans Tied to Fewer Behavioral Symptoms in Dementia

A small randomized trial of 36 caregiver-dyads in South Korea found that a mobile app translating caregiver-logged behavior into individualized, nonpharmacological activity suggestions was associated with lower overall behavioral and psychological symptoms, less agitation and depression, better nighttime sleep efficiency, and higher caregiver confidence, researchers reported.

The pill that tells you what it is

A methods paper in PAIN, posted online ahead of print on August 10, 2026, sets out for the first time how to run placebo trials in which nobody is fooled. It is written for pain research, but the design questions it raises reach into every trial that has a placebo arm, dementia trials included.

The Shock That Hurt More Because It Was Dreaded

In a Taipei scanner, 31 healthy adults were given the same electric jolt over and over. What changed how much it hurt was not how reliable their expectation was, but how much feeling it carried. For anyone thinking about pain and dementia, that difference matters.

The Alzheimer's test that arrives in the post

A multicenter European study put a finger prick against a needle in the arm, and asked whether a card of dried blood mailed at room temperature could carry the same signal. For people whose veins are hard to find, including adults with Down syndrome, the question is not academic.

The interviewer with all the time in the world

At a cognitive neurology clinic, a voice AI ran the long history-taking interview while patients waited for their testing. A peer-reviewed SIGDIAL 2026 paper reports what it caught, and what it still cannot tell anyone.

The Wrist Knows Something the Scanner Confirms Later

A study of 344 cognitively healthy adults in the Baltimore Longitudinal Study of Aging paired seven days of wrist actigraphy with brain MRI, and found the link between a ragged 24-hour rhythm and shrinking brain tissue was clearest in the oldest participants. It is observational, the follow-up was short, and it diagnoses nobody.

Spinal cord injury drug eases DNA damage in mouse model of Alzheimer's

A small molecule that already cleared human safety testing for spinal cord injury reduced neuronal DNA damage and brain inflammation in Alzheimer's mice, King's College London researchers report, though the study involved only a handful of animals and has not yet been tested in Alzheimer's patients.

Small Study Ties Gut-Linked Blood Metabolites to Alzheimer's Diagnosis

A case-control study combining plasma chemistry with gut bacteria data in 85 people found a three-metabolite blood panel that distinguished Alzheimer's patients from cognitively normal controls, and flagged a cell-energy pathway the authors say deserves closer study, though they call the tool an aid, not a stand-alone diagnostic.

The brain keeps two ledgers on pain: how bad, and how sure

A PLOS Biology study of 72 adults, published on 2 March 2026, separated the strength of a pain expectation from its certainty, and found each one riding a different alpha rhythm in the anticipating brain. Basic mechanism work in healthy volunteers, not a treatment, and not a dementia study.

The Care Plan That Writes Itself, In A Demo

A preprint from four University of Missouri researchers describes AgenticAD, a system of eight AI agents in which three of them interview a caregiver and produce an assessment summary, a daily plan and a follow-up schedule. It has been run on three invented personas and no real families.

The Heart Signal That Failed Alone Gets a Second Reading

Months after a Dutch cohort found heart rate variability predicted almost nothing over 13 years, a peer-reviewed narrative review argues the metric was never meant to be read by itself. Its proposal: pair it with the brain's own electrical rhythm.

Fifty-Eight Studies In, One Bacterium They Agree On

A scoping review in Alzheimer's & Dementia sifted 4,751 papers on the microbiota-gut-brain axis in mild cognitive impairment and Alzheimer's. What survived the sift says as much about the field's habits as about anyone's gut.

Brain scans show two routes to placebo pain relief are not the same

A pooled analysis of brain-scan data from 409 people finds that words alone and words paired with conditioning change pain-related brain activity in overlapping but partly different ways, a distinction researchers say matters for how future placebo-based protocols are designed and tested.

The cells that stopped dividing and started shouting

A new synthesis from Barcelona and London argues that brain aging is driven not only by misfolded protein, but by cells that have quit the cell cycle and will not stop signalling. It is a review, not a discovery, and the human evidence is still thin.

The leaflet, the search bar, and the thing that changed

Three pre-registered experiments from the University of Toledo found that explaining the nocebo effect did not just shift how people felt about side effects. It changed how much side-effect information they went looking for, and it did not leave them more anxious.

Wrist Sensors Trace Circadian Timing to Measured Tau, Not Just Diagnosis

A month of round-the-clock wrist accelerometry in 68 older adults with early cognitive impairment found that earlier, not later, timing of peak daily activity tracked with higher amyloid-beta and tau on PET scans, and that fragmented daily rhythms tracked with tau in later-stage brain regions. The study is cross-sectional and cannot say which came first.

Twenty years between the question and the blood test

A handful of yes-or-no questions about sleep, answered by 1,325 Americans at an average age of 40, tracked with Alzheimer's blood markers measured two decades later. The study pushes the sleep question earlier than it has ever been asked, and shows how thin our measure of sleep still is.