Treat Test May Spot Dog Dementia Signs Years Before Owners Notice

Three Treat Games Measure Memory, Self-Control, and Flexibility in Dogs

A dog’s thinking skills can start showing measurable signs of aging around age 8, potentially years before an owner notices anything wrong, according to new research testing 239 companion dogs. The tool behind the finding takes about 40 to 60 minutes and relies on three simple food-based puzzles.

University of Arizona researchers working with the Dog Aging Project built the exam, called CANID. A senior dog that gets lost at home or grows more anxious may be showing signs of dementia, but by then the decline may have been underway for years. Most existing tools are owner surveys, and lead author Stephanie Hargrave said that by the time they identify a dog as having dementia, the impairment “is often quite severe.” Dogs make useful stand-ins for human dementia research, since some develop brain changes resembling features of Alzheimer’s disease.

Published in the journal Animal Cognition, the study brought the same dogs back six months and a year later to see whether their scores held steady.

Three Food Puzzles Make Up the Dog Cognition Test

CANID grew out of an earlier version that used five puzzles across two sessions of 30 to 60 minutes each. That version found older dogs performing noticeably worse on memory and self-control tasks, but it took too long for large groups of pet dogs. Researchers kept the three puzzles most strongly tied to age and fit them into one visit. Senior author Evan MacLean explained the problem: “We can’t ask them questions; we can’t tell them, ‘OK, this is the test of your memory.'” So the dogs play games for treats instead.

Spatial Reversal tests how well a dog adapts when rules change. A treat sits behind a panel with a barrier blocking one side path, and the dog learns to use the open one. Then researchers move the barrier, and the dog must try the opposite side. Delayed Search is hide-and-seek for treats, a short-term memory check: a dog watches a treat disappear behind one of four screens, then waits up to 40 seconds, sometimes with a distraction, before searching. The Cylinder task measures impulse control. A treat sits inside a clear tube open at one end, and the dog has to resist lunging at the visible food and walk around to the opening.

Dog Cognition Test Scores Dropped as Dogs Aged

Researchers tested 239 dogs, from under a year old to about 16, averaging 7 years old. They came from a university lab and a dog daycare in Tucson, Arizona, and a service dog training center in Santa Rosa, California. Not every dog came back: 138 returned at six months, 114 at one year, and 91 completed all three rounds.

Age accounted for about 27% of the differences in scores between dogs, and scores fell with age on all three puzzles. According to the authors, questionnaire-based assessments do not show a sharp rise in the chance of a dementia diagnosis until roughly age 10 to 12, a few years after differences first appear in CANID scores.

Dogs from the service dog center, 31 of them, scored higher than dogs at the other sites after accounting for age, mostly on Cylinder and Spatial Reversal, the two puzzles that test restraint. Training history, temperament, and the genetic similarity of dogs bred for assistance work are possible explanations.

Retest Scores Stayed Consistent Despite a Practice Boost

Dogs that scored high once tended to score high again at six and twelve months, a quality called repeatability, though part of that consistency comes from age, since older dogs score lower at every visit.

At six months, overall scores rose modestly, the same kind of practice boost seen when people retake memory tests and about the same size as the average gain in a large review of human cognitive tests. By twelve months, scores had settled back near where they started. Individual puzzles varied: impulse-control gains lasted through both retests, rule-switching improved and then faded, and hidden-treat memory at one year fell below the starting point. The authors say that dip may reflect real decline. Overall scores did not fall, and the researchers offer two possible reasons why: many dogs in the sample were young, and practice gains can cancel out real decline.

Dog Test Has Not Yet Shown It Can Predict Dementia

Because the study did not include owner questionnaires or biological samples, it remains unknown how closely CANID scores match other measures of canine cognitive decline, whether they predict which dogs will develop dementia, or whether the test can separate normal aging from early disease. Testing also requires food-motivated, calm dogs, trained staff, and specialized equipment, so owners cannot run it at home.

For now, CANID is a research tool with promise, not a diagnosis. For Hargrave, whose own Chihuahua mix, Roxy, is getting up in years, the goal is personal: “I’m especially interested in what I can do to prolong her cognitive health and prevent decline.” Whether a treat test can eventually give owners that head start is the question future studies have to answer.

---