Canine Cognition Study Uses Games to Evaluate Best Candidates for Service Dogs
Only about half of dogs trained to be service dogs make it. Considering the investment of time and money to train a dog selected for service work, coupled with the high demand for service dogs to help people with physical and developmental disabilities, the ability to predict which dogs are more likely to succeed would potentially enable more dogs to be trained and placed with people who need them.
That’s the aim of a collaborative two-year study involving researchers at the Duke Canine Cognition Center in Durham, North Carolina, and the largest nonprofit assistance-dog training organization in the country, Canine Companions for Independence in Santa Rosa, California. Funded by the AKC Canine Health Foundation, the study will produce a series of cognitive tests, or games, to help assess canine emotional health and well-being in the service dog selection process. The research also might help predict specific types of service for which a dog is best-suited.
“Until now, there hasn’t been any attempt to develop cognitive testing for service dog assessment,” says Evan MacLean, PhD, co-director of the Duke Canine Cognition Center. “Are there individual differences in cognition that relate to how well dogs fare in their working lives? There may be different cognitive styles in dogs that align with their potential to work in various capacities.”
The collaboration has allowed the canine cognition experts to tap into information about what makes a service dog successful and gain insights about their behavior and training. Assistance dogs trained by Canine Companions are helping to develop the cognitive tests by being the first to participate.
The long-term goal of the research is to streamline the process of breeding, training and matching dogs with jobs and handlers. “Does a dog’s communication ability, memory, empathy for humans, and ability to independently solve problems matter?” asks MacLean. “We believe that temperament is important, but so is a dog’s cognitive aptitude in determining the best service dogs.”
Although there has not been a lot of research looking at psychological differences among dogs, this study will attempt to understand the nature of individual dog behaviors. “We are trying to understand if dogs vary in the same ways that people do,” MacLean says. “Are there similar cognitive profiles in dogs?”
About 200 dogs will be included in each of the two phases of the research. The large number of dogs allows the researchers to evaluate whether there are clusters of skills that reliably group dogs together and whether this brings new understanding about dog psychology.
A shortage of service dogs to assist people with disabilities is universal. Jeanine Konopelski, director of marketing for Canine Companions, says, “As people have learned about the life-transforming benefits an assistance dog can provide a person with a disability, applications have been on the rise. Canine Companions has over 400 people currently waiting to receive a highly trained assistance dog.”
Trained assistance dogs can alert people to sounds and help them balance, pick up dropped objects, open doors, and push buttons. Service dogs can pull wheelchairs and alert authorities and family members to medical conditions such as hypoglycemia and seizures. Dogs provide comfort, confidence and focus to people suffering from post-traumatic stress disorder and to autistic children.
“It costs up to $45,000 to train a service dog, which includes up to two years of training. It begins with initial obedience commands, socialization and training by a professional dog instructor and continues for the life of the dog with ongoing follow-up,” Konopelski says.
Sadly, 50 percent of dogs in training never graduate due to a variety of behavioral or health issues. “It takes an extremely special dog to become a successful assistance dog,” says Konopelski. “It takes one that is not only physically sound but also secure in all environments, eager to work yet has a moderate level of energy, and able to ignore distractions such as birds, people and other dogs.”
To increase the odds of successfully selecting dogs for training, many service dog organizations breed their own dogs. Canine Companions is one such organization. Golden Retrievers, Labrador Retrievers and Golden-Labrador crossbred dogs are used in its program.
Puppies are whelped in the homes of volunteer breeders with whom the dam lives. When puppies are 2 months old, they are examined and vaccinated by a veterinarian at Canine Companions and then placed in the homes of volunteer puppy raisers. At 14 months of age, they return to Canine Companions to begin formal assistance dog training, which lasts from six to nine months.
The screening process starts with temperament observations that are noted by the puppy raiser. Standardized temperament testing is used when dogs are around 14 months of age before formal training begins. The most widely used test is the Canine Behavioral Assessment and Research Questionnaire (C-BARQ). The families who have raised the dog complete the 36-question C-BARQ test, which is intended to evaluate behaviors such as aggression, excitability, dependence, and fearfulness.
A recent study of 7,969 assistance dogs representing five service dog organizations looked at the effectiveness of the C-BARQ questionnaire. The dogs were assessed at 6 and 12 months of age. Of the 36 questions, 27 were considered to have a predictive value. The most predictive question was whether the dog “pulls excessively hard on the leash.” Dogs more likely to pull hard were less likely to succeed.
The nearly completed first phase of the Duke study entailed using a large battery of tests, or games, to measure 28 problem-solving skills based on previously published animal cognition studies. Each of 200 dogs were tested in four one-hour sessions. The goal is to shorten the test and use only the most effective games as quality measures of cognition.
“We started with a very large battery because we don’t know which types of measures may be related to success as an assistance dog,” McLean says. “These tests cover diverse aspects of problem-solving that include impulse control, memory, understanding human communication, and inferential reasoning.”
The second phase, a shorter version of the phase-one games, will include a different set of 200 dogs. The shorter test will feature measures that have shown promise in discriminating between dogs that succeed or fail in training. This phase will begin later this year and is expected to take about one year to complete.
“The cognitive traits needed by dogs that guide the blind may be different than those needed by dogs that alert deaf people to sounds or those that help people in wheelchairs,” MacLean explains. “We may be able to gather information from dogs to help us steer them toward the ‘careers’ that best suit their dispositions.”
Pet dogs also can participate in the phase-two testing. They will provide a benchmark that can be used to determine whether the cognitive traits seen in assistance dogs are typical of dogs in general or the result of selective breeding and training. Owners may enroll their dogs by going to dukedogs.com.
Hopefully, the results of this study will produce cognition testing that can be used to help meet the ongoing need for service dogs to help people who need them. “By working with Canine Companions, we have gained firsthand exposure to how assistance dogs can make a huge difference in peoples’ lives,” MacLean says. “That’s powerful motivation to do the best we can to help streamline the process of breeding, training and matching dogs with jobs and handlers.”
Using Cognitive Test Games to Predict Ability
For the first time, a series of cognitive games are being used to predict the abilities that best define important characteristics for assistance dogs. Here are examples of the cognitive tests being developed by experts at the Duke Canine Cognition Center.
Which Cup? A treat is hidden underneath one of two inverted cups. The handler stands between the cups and points to one cup. Will the dog interpret the pointing correctly and go to the cup with the treat? What if the handler simply looks at the correct cup or places a marker on the cup. Many dogs, but not all, are able to use these cues to find the treat.
Loose Lid? A dog is given a simple task, such as removing a loose lid placed over a container with a toy. Most dogs find this task easy to solve, but what if the lid is secured tightly to the container so it cannot be opened? Some “do-it-yourself” dogs continue to work at the lid, while “help seekers” give up easily and solicit help form the handler.
FIDO Technology Equips Service Dogs for Success
A Canine Companions for Independence service dog, an 8-year-old black Labrador-Golden Retriever crossbred named “Caspin,” is using state-of-the-art technology to alert for help when his handler, Wallis Brozman of Santa Rosa, California, needs it. A vest worn by the dog has sensors that activate a prerecorded message announcing that Caspin’s handler needs attention.
Brozman has a condition that limits her ability to speak louder than a whisper. When she needs help, she can rely on Caspin to find the nearest person and tug on a sensor on the vest. Caspin is trained to return to Brozman’s side with help.
The technology, called Facilitating Interactions for Dogs with Occupations (FIDO), was developed by researchers at the Center for BioInterface Research at Georgia Tech in Atlanta partnering with Canine Companions for Independence. The National Science Foundation is funding the research.
Melody Jackson, PhD, a five-time puppy raiser for Canine Companions and a member of the Canine Companions Research Committee, headed the project. “We are testing the FIDO technologies with service dog teams and service dogs in training,” she says. “The idea is to develop devices that can help service dogs communicate problems their handlers may be having by biting, tugging or touching them with his or her nose. Some vests are even equipped with GPS to help authorities locate someone needing help.”