News · Brain & Mental Health
Mood that swings predicted dementia better than mood that is low
Following 11,951 older adults for up to 19 years, researchers found that fluctuating depressive symptoms predicted later cognitive impairment better than how depressed someone was at any single visit.
Based on 11,951 older adults across 42 US Alzheimer's Disease Research Centers
- 11,951 older adults across 42 US research centers, followed for up to 19 years.
- Fluctuating mood scores tracked with more new cases of impairment and dementia.
- Instability predicted better than baseline severity, which is what models usually use.
- Adjusted for the APOE e4 gene variant, the main inherited Alzheimer's risk factor.
- Observational. Early brain disease can itself destabilize mood, so direction is unclear.
Depression in later life is a known risk factor for dementia. Risk models capture it the way they capture most things: with a single number, taken once, describing how depressed someone was at the start.
A study in Alzheimer’s & Dementia argues that the number is the wrong shape. What predicted better was not how low a person’s mood sat, but how much it moved.
What was done
Data were drawn from 11,951 older adults enrolled across 42 US Alzheimer’s Disease Research Centers and followed for up to 19 years.
Depressive symptoms were assessed using the Geriatric Depression Scale, a standard questionnaire for older adults, administered repeatedly rather than once. That repetition is the whole design. Instability cannot be seen in a single measurement, which is why most models have never looked for it.
The researchers then compared several competing accounts of how mood over time might relate to later diagnosis, and adjusted for the APOE e4 gene variant, the most common inherited risk factor for Alzheimer’s disease.
What they found
Fluctuating depressive symptom trajectories were associated with increased risk of incident cognitive impairment, meaning new cases arising during follow-up rather than cases already present.
More pointedly, symptom instability showed the lowest prediction error among the accounts tested. Instability beat severity at its own job.
The authors’ reading is that instability in depressive symptoms may represent an early signal, appearing before the cognitive diagnosis rather than causing it.
Why it might work that way
Two explanations sit behind a finding like this, and the study does not choose between them.
The first is that a mood that will not hold steady reflects a brain already under strain. The regions that regulate emotion overlap with those affected early in several dementias, and disease can destabilize mood long before it disturbs memory.
The second is that instability is a marker of something else entirely, such as vascular disease, poor sleep, medication changes or the accumulating losses of later life, each of which independently tracks with cognitive decline.
Either way, the practical implication is the same: a mood that swings deserves more attention than a mood that is reliably a little low, which inverts the usual clinical instinct.
What this does not settle
Direction, above all. This is observational, and the alternative story, that early disease causes the mood instability, fits the data just as well as the story where instability comes first.
The population is also a specific one. Alzheimer’s Disease Research Centers enroll people who volunteer for dementia research, many with a family history or existing concerns about memory. They are not a cross-section of older adults, and rates of both mood symptoms and diagnosis will differ.
The instability measure itself is new enough that it has no agreed definition. Different rules for what counts as fluctuating will produce different results, and the paper’s own analysis of this is described as exploratory.
What to take from it
Not a self-test. The finding lives at the level of thousands of people over nearly two decades, and no individual should read their own mood chart against it.
What it changes is what researchers should collect. A risk model that asks about depression once is throwing away the part of the signal that predicted best, and asking twice costs almost nothing.
People also ask
What does symptom instability mean here?
Not how low someone's mood got, but how much it moved between visits. A person scoring moderately depressed at every assessment is stable. A person swinging between clear symptoms and none is unstable, even if their average is lower. Risk models almost always use the average or the worst reading, which is precisely the information this study says is less useful.
Does this mean depression causes dementia?
It cannot show that, and the reverse is a live possibility. The brain changes of Alzheimer's disease begin years before memory symptoms and affect the regions that regulate mood, so unstable mood in someone's seventies may be an early manifestation of the disease rather than a cause of it. The authors describe instability as potentially an early marker, which is a claim about timing rather than causation.
How was mood measured?
With the Geriatric Depression Scale, a standard questionnaire designed for older adults, administered repeatedly over years. Repeated measurement is what makes the study possible: instability is invisible if you only ask once, which is how most risk models are built.
What is APOE e4 and why adjust for it?
It is the most common inherited risk variant for Alzheimer's disease, carried by roughly a quarter of people. Adjusting for it means the finding is not simply genetic risk expressing itself as both mood changes and dementia. That strengthens the result without settling what is driving it.
Should someone with fluctuating mood be worried?
This is general information, not a diagnostic tool, and mood fluctuates for many ordinary reasons including bereavement, illness, medication and life circumstance. The finding is about patterns across a large group over many years, not a test that can be applied to one person. Anyone concerned about their memory or their mood should raise it with a doctor rather than self-assess.