Can More Walks Prevent Dog Dementia? Reading the 15,019-Dog Study

Last updated: August 15, 2026

Not as far as this study goes. Dogs that move less do have more dementia, but the researchers who ran the survey added a line to the paper: the same numbers appear if the order runs the other way — the dog stopped moving because cognition declined.

Of 15,019 owner responses, 1.4% of dogs met the dementia threshold
Holding other factors equal, each additional year of age multiplied the odds by 1.52
Dogs described as barely active had 6.47 times the odds of very active dogs
The design cannot establish which came first. Nowhere does the paper say exercise prevents it

Read about night howling or wandering in old dogs and you keep meeting the same explanation: too little exercise makes dementia more likely. Follow it back and it usually leads to one large American survey. We read the paper behind the 15,019 owner responses.

The short version: this survey never says exercise prevents anything. There is a sentence in the text placed specifically to head off that reading.

What gets said

"More walks and more play will prevent dementia in your dog." It comes up again and again in writing for owners. And the survey does show a strong association between activity and dementia. Looking only at the number, the temptation is understandable.

A related claim is that dementia is not rare, or that some share of senior dogs have it. The same survey often gets cited for that too, and that is where the number needs careful handling.

How the survey was built

Owners taking part in a large American long-term research project answered a questionnaire. The sample was 15,019 dogs, mean age 6.9 years with a standard deviation of 4.2 years. Young dogs through old dogs, in other words — a wide spread.

The important part: these are not veterinary diagnoses. Owners answered questions about behavior, and dogs whose scores crossed a line were classified as meeting the criteria for dementia. Where to draw that line was taken largely as-is from earlier researchers, and the paper lists this as a limitation.

A cross-sectional study looks at the state of things at a single point in time. That is different from following the same dogs for years to see what became of the ones that exercised (a longitudinal study). A cross-sectional design can show that two things occur together, but not which came first.

So this survey did not compare active and inactive dogs over years. It looked at both groups once, at the same moment. The scale is large; the timeline is not there.

The numbers

1.4% met the threshold

Of 15,019 dogs, 1.4% met the criteria for dementia. That may read as lower than expected. The figure includes puppies and young adults; it is an average across every age.

The paper says as much: this 1.4% lines up well with previously reported rates of veterinary-diagnosed dementia, while studies that recruited older dogs have reported higher rates. It is 1.4% of all dogs, not 1.4% of senior dogs.

1.52 times per year of age

Holding age, health status, breed type, and neuter status equal, each additional year of age multiplied the odds of meeting the criteria by 1.52.

An odds ratio of 1.52 can be read, roughly, as "the likelihood goes up about 1.5 times with each year." Because that multiplies year on year, the increase gets steeper as the dog gets older.

A 12-year-old and a 13-year-old are one year apart, but on this estimate the gap is 1.5 times. At 14 it is steeper again. The overall figure looked low at 1.4% because most of the dogs in the sample were young.

Dogs that barely move: 6.47 times

Comparing dogs of the same age, health status, breed type, and neuter status, those described as "barely active" had 6.47 times the odds of those described as "very active." That is a large gap, and age does not explain it.

Confidence intervals and sample breakdown

Sample: 15,019 dogs (mean age 6.9 years, standard deviation 4.2 years)
Met dementia criteria: 1.4% (the paper gives no headcount; the percentage works out to roughly 210 dogs)
Odds ratio for age: 1.52 (95% confidence interval 1.44–1.61)
Odds ratio for activity: 6.47 (95% confidence interval 2.93–17.23)

The width of that activity interval matters. Rather than a single figure of 6.47, it is more accurate to hold it as somewhere between about 3 and 17 times. With only 1.4% of the sample meeting the criteria, the estimate gets less stable the more finely the data is divided.

What is not known

A figure like 6.47 makes you want to get the leash. The paper, though, attaches this to it:

…it is important to note that this correlation could also exist simply because of dogs exhibiting less activity due to their cognitive decline. Additionally, the presence of a third, unmeasured variable (such as owner interaction) could be amplifying this association.
From the paper

In other words, this survey cannot separate "too little exercise led to dementia" from "dementia led to less movement." The authors also raise the possibility that unmeasured factors, such as how much the owner engages with the dog, are strengthening the association.

Dogs with dementia may wander without purpose, but they are also known to spend more of the day dull and still. If that is what is happening, "not moving" is the symptom, not the cause. The same number supports an explanation with cause and effect swapped.

One more thing: every data point here comes from owners. The paper lists this as a limitation: "Owner-reported information is potentially susceptible to multiple forms of bias." An owner who already suspects dementia will recall behavior changes in more detail.

Drawing a line on a behavior questionnaire has its own limits too. The paper reports high predictive capacity for the threshold, while noting that it "was chosen by previous researchers, and measures of sensitivity and specificity would be difficult to obtain."

What this survey is good for

As a guide to age

What this survey does establish is the relationship with age. An increase of about 1.5 times per year is a usable guide for when to start paying attention.

A separate study reported age-related behavior changes in roughly 30% of dogs at 11–12 years and roughly 70% at 15–16 years. That is the share showing behavior changes, not the share diagnosed with dementia — a different measure. But it points the same direction: the numbers climb from around age 12.

As something to bring to the vet

Reduced movement is worth reporting on its own. Whether the cause is joints, eyes and ears, or the brain is not something you can separate at home. Adding "she has not wanted to walk much lately" or "he comes when called less often now" gives your vet more places to look.

Say when it started and how often it happens, and it becomes concrete: "for about two months, she has woken and cried two or three nights a week."

Not as a reason to blame yourself

This research is not grounds for concluding that too few walks caused your dog's dementia. The researchers wrote plainly that they cannot tell the order. A number with no established sequence cannot locate the cause in you.

If there is something to do, it is less about exercise itself and more about the shape of the day: more stimulation during daylight, quiet after dusk. That structure shows up as standard advice for night howling in veterinary sources too. The practical side of care is on the daily care page.

How this number gets misread

Misreading 1: "1.4% — so dementia in dogs is rare"

That 1.4% averages puppies together with old dogs. Given an increase of about 1.5 times per year, the figure for senior dogs alone is far higher. The paper itself notes that studies recruiting older dogs report higher rates.

Reading "1.4% of all dogs" as "1.4% of senior dogs" is the most common error here. It errs in the reassuring direction, which is why it slips past.

Misreading 2: "6.47 times — so exercise is preventive"

The paper states outright that the same correlation appears if cognitive decline is what reduced the movement. This is not a study showing a preventive effect.

Testing whether exercise prevents anything would require following dogs with different activity levels for years and comparing who developed it. This survey is not built that way.

Misreading 3: "6.47 is itself a solid number"

That 6.47 carries a range from about 3 to 17. With only 1.4% of the sample meeting the criteria, the estimate wobbles the more finely the data is split. Treating "6.47" as meaningful to two decimal places goes beyond what this study can support.

Articles describing the research almost always drop the range. The sentence still works without it, but the sense of how firm the finding is does not survive.

Common questions

How common is dementia in dogs?

In this survey, 1.4% of 15,019 dogs met the threshold (the paper reports the percentage only, not a headcount). But that is an average across puppies and old dogs together. The odds rise about 1.5 times with each additional year of age, so the figure for senior dogs alone is far higher. The paper itself notes that studies recruiting older dogs have reported higher rates.

Can exercise prevent dementia?

This survey cannot answer that. Dogs described as barely active had 6.47 times the odds of very active dogs, which is a strong association. But the authors add in the paper that the same correlation would appear if declining cognitive function were simply reducing the dogs’ activity. The study design cannot separate cause from effect.

At what age should I start watching for this?

The survey does not name a cutoff age. What it shows is the rate of increase: about 1.5 times per year. A separate study reported age-related behavior changes in roughly 30% of dogs at 11–12 years and roughly 70% at 15–16 years. Around age 12 is a reasonable point to stop filing restlessness at night, or stopping in a corner of the room, under "just getting old."

Was the 1.4% diagnosed by veterinarians?

No. Owners answered a questionnaire about behavior, and dogs whose scores crossed a set threshold were counted. That threshold was taken largely as-is from earlier researchers, and the paper lists this among its limitations.

My dog has stopped moving around. Is it dementia?

This survey cannot tell you. Reduced movement has many causes that are impossible to separate at home — joint pain, changes in sight or hearing, internal illness. The change itself is worth reporting at a visit, so try saying something specific like "she has not wanted to walk much lately."

Sources

Yarborough S, Fitzpatrick A, Schwartz SM, et al. Evaluation of cognitive function in the Dog Aging Project: associations with baseline canine characteristics. Scientific Reports. 2022;12:13316.doi:10.1038/s41598-022-15837-9(accessed August 14, 2026)

The figures of roughly 30% at 11–12 years and roughly 70% at 15–16 years come from a separate study (from the care-article research verified on August 9, 2026).