40

SHARES

Share on Facebook

Post on X

A dog remembers where food was hidden, switches routes when a familiar path is blocked, or works around a clear barrier instead of reaching straight toward a treat. A newly published assessment combines these kinds of tasks to help researchers study how dogs' cognitive performance differs with age.^[1]

The assessment is called CANID, short for Cognitive Assessment for Neuropsychological Impairments in Dogs. In a study published October 12 in Animal Cognition, Stephanie Hargrave and colleagues describe a three-task battery designed for companion dogs, with repeat testing to examine how consistently it measures performance.^[1]

The important distinction is what the study establishes: CANID detected age-related differences in this sample, but it has not yet been shown to diagnose dementia or reliably measure decline in an individual dog.^[1]

Three tasks, three different challenges

The researchers describe sessions lasting approximately 40 to 60 minutes, with a brief vision and hearing screen alongside the cognitive tasks.^[1]

• Changing a learned response. In the spatial reversal task, a dog learns which side of a panel provides access to food. The open route then switches, requiring the dog to change its previously successful approach.^[1]
• Remembering a location. In delayed search, the dog watches food being hidden in one of four locations, then searches immediately or after a delay. Some trials also include a distraction during the wait.^[1]
• Working around a visible barrier. In the cylinder task, the dog first learns to retrieve food through the open end of an opaque cylinder. When the cylinder becomes transparent, the challenge is to use that opening rather than make direct contact with the barrier toward the visible food.^[1]

Together, the tasks sample memory and executive function, the processes involved in adjusting behavior and inhibiting an immediate response.^[1]

These are standardized research procedures, not three casual games with a diagnostic score attached. The study used trained experimenters, a handler, detailed scoring rules and video recordings, with a second observer checking a subset of trials.^[1]

What the researchers found

At baseline, 239 dogs contributed data from at least one task. They ranged from approximately one to 16 years old and were tested at a university cognition center, a dog daycare facility and a service-dog training center in the United States.^[1]

Overall CANID scores were generated for 233 dogs at baseline, with some incorporating an estimated score for a missing task. The paper reports usable follow-up data from 138 dogs at six months and 114 at twelve months.^[1]

The dataset also included dogs tested with an earlier, longer battery. Their results were restricted and rescored to match the shorter assessment, so not every record came from the new single-session format.^[1]

Older age was associated with lower overall scores, and the authors describe age-related differences becoming apparent around eight years in this sample. That is a research finding across dogs, not an age at which every dog becomes impaired.^[1]

Scores showed consistency across repeat assessments, although some of that consistency reflected the strong association with age. After accounting for age and other modeled factors, repeatability of the overall score was moderate.^[1]

The retests also revealed why cognitive aging is difficult to measure. Overall scores improved modestly at six months but were not significantly different from baseline at twelve months. The individual tasks behaved differently: cylinder performance remained improved at twelve months, while delayed-search performance was lower.^[1]

The authors caution that familiarity with a task and age-related decline can occur together. A stable combined score could therefore conceal opposing changes, rather than prove that nothing has changed.^[1]

Why the findings matter to trainers

One useful feature is the separation of different demands. Remembering a location is not the same challenge as abandoning a previously rewarded route or resisting a direct approach to visible food.^[1]

For trainers, that offers a more precise way to think about a difficulty than simply calling a dog “less responsive.” It encourages a question: what, exactly, is this situation asking the dog to do? That is a practical interpretation of the study, not a training method the researchers tested.

The testing context also matters. Dogs needed sufficient food motivation and comfort with the setting and unfamiliar people. Performance differed between testing sites, but the study cannot establish whether those differences resulted from training, temperament, genetics, the environment or a combination of factors.^[1]

It would therefore be a mistake to treat the better scores at the service-dog center as proof that a particular training program prevents cognitive aging.^[1]

A research tool, not a home diagnosis

The study did not compare CANID scores with cognitive-dysfunction questionnaires or biological samples. The authors identify further work needed to distinguish normal aging from pathological decline, follow individual dogs over longer periods and establish how much score change exceeds measurement error.^[1]

That leaves CANID with a useful but bounded role: a standardized, openly documented approach for studying cognitive aging, with evidence on age differences and repeatability. It is not a substitute for veterinary assessment, and this study does not show that performing the tasks prevents dementia.^[1]

Readers can examine the original abstract and download the complete, unmodified paper in our Knowledge Base. The resource also links to the publisher's supplementary material and any later updates.^[1]

Sources

[1] Hargrave, S. H., and colleagues (2026). Cognitive assessment for neuropsychological impairments in dogs (CANID): original abstract and full paper. Animal Cognition. Published October 12, 2026. DOI: 10.1007/s10071-026-02112-x. Our Knowledge Base includes the unmodified publisher PDF and the original source link.