research paper about dogs

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

Depression, anxiety, and happiness in dog owners and potential dog owners during the COVID-19 pandemic in the United States

Roles Conceptualization, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Writing – original draft, Writing – review & editing

* E-mail: [email protected]

Affiliation Nestlé Purina Research, Saint-Louis, MO, United States of America

Roles Methodology, Validation, Writing – original draft

Roles Writing – original draft, Writing – review & editing

Roles Data curation, Formal analysis, Writing – review & editing

Roles Data curation, Formal analysis

Roles Conceptualization, Writing – review & editing

• Francois Martin, 
• Katherine E. Bachert, 
• LeAnn Snow, 
• Hsiao-Wei Tu, 
• Julien Belahbib, 
• Sandra A. Lyn

• Published: December 15, 2021
• https://doi.org/10.1371/journal.pone.0260676
• Reader Comments

Major life events, such as the COVID-19 pandemic, affect psychological and physiological health. Social support, or the lack thereof, can modulate these effects. The context of the COVID-19 pandemic offered a unique opportunity to better understand how dogs may provide social support for their owners and buffer heightened symptoms of stress, anxiety and depression and contribute to happiness during a major global crisis. Participants (768 pet dog owners and 767 potential pet dog owners) answered an online survey, including validated depression, anxiety, happiness psychometric scales, attitude to and commitment towards pet, and perceived social support. Potential pet dog owners were defined as individuals who did not own a dog at the time of the survey but would be very or extremely interested in owning one in the future. Dog owners reported having significantly more social support available to them compared to potential dog owners, and their depression scores were also lower, compared to potential dog owners. There were no differences in anxiety and happiness scores between the two groups. Dog owners had a significantly more positive attitude towards and commitment to pets. Taken together, our results suggest that dog ownership may have provided people with a stronger sense of social support, which in turn may have helped buffer some of the negative psychological impacts caused by the COVID-19 pandemic.

Citation: Martin F, Bachert KE, Snow L, Tu H-W, Belahbib J, Lyn SA (2021) Depression, anxiety, and happiness in dog owners and potential dog owners during the COVID-19 pandemic in the United States. PLoS ONE 16(12): e0260676. https://doi.org/10.1371/journal.pone.0260676

Editor: Simon Clegg, University of Lincoln, UNITED KINGDOM

Received: August 30, 2021; Accepted: November 12, 2021; Published: December 15, 2021

Copyright: © 2021 Martin et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: The authors received no specific funding for this work.

Competing interests: The authors declared that no competing interests exist.

Introduction

Worldwide, the socio-economic impacts of the COVID-19 pandemic have been extensive. Governments issued quarantine and social distancing policies as well as lockdown measures to mitigate the transmission of the disease. In the United States, there was no unified or enforceable federal response, but the federal government issued a series of public health recommendations aimed at preventing the transmission of SARS-CoV-2 (e.g., social distancing, masks wearing, temporary voluntary national lockdown in the spring of 2020). Unemployment and decreased consumer spending contributed to economic recessions. Such socio-economic factors have led to higher levels of loneliness and social impairment [ 1 , 2 ]. The prolonged and significant disruption in the daily lives of people caused by the pandemic increased stress, anxiety, loneliness, and depression levels in many [ 3 – 19 ].

Major life events, such as the COVID-19 pandemic, affect psychological and physiological health. Social support, or the lack thereof, can modulate these effects. Social support is defined as the outcome of one or more of the following: 1) the feeling of being cared for; 2) the belief that one is loved, esteemed, and valued, and 3) the sense of belonging to a reciprocal network [ 20 , 21 ]. Within this framework resides the concept that social support provides protection from pathological states and accelerates recovery from illness by acting as a buffer in times of crisis [ 22 , 23 ]. Empirical studies that explored people’s social support found a positive link between adequate social support and mental health outcomes [ 24 – 32 ]. In turn, lacking social support has been associated with negative impacts on people’s well-being [ 33 – 36 ].

Applying the concept of social support to human-animal relationships is a logical extension. When describing the advantages of pet ownership, people will often mention the emotional support and self-esteem gained from the relationship [ 37 ]. Pets provide emotional stability and affection during stressful events such as divorce or bereavement [ 38 , 39 ]. Pets are perceived as always available, predictable in their responses, and non-judgmental [ 40 – 43 ]. In addition, they are considered to be dependent and caring towards their owners with unconditional love [ 44 , 45 ]. Pets provide tactile comfort [ 46 ] and recreational distraction from worries [ 47 , 48 ]. In contrast with other social interactions, no special social skills are usually required to elicit a positive response from a pet [ 49 ]. Pets’ reactions are not based on who the person is or their social competence, which provides a level of ease and sense of relief not usually experienced in other human-human social interactions [ 44 , 50 ].

Prior evidence suggests that pet dogs in particular may provide social support to humans by contributing to enhanced positive affective states and decreased sadness, anxiety, and loneliness [ 51 – 63 ]. However, not all studies have reported positive effects of dogs on their owners’ well-being [ 64 – 66 ]. In fact, some studies found that dog ownership may increase an owner’s stress levels [ 44 , 67 – 71 ]. The mixed results indicated by research to date reveals a gap in knowledge regarding when and how pet dog ownership contributes to greater wellbeing among pet owners.

The context of the COVID-19 pandemic offered a unique opportunity to better understand how dogs may provide social support for their owners and buffer heightened symptoms of stress, anxiety and depression and contribute to happiness. Recently published studies suggest that pet ownership during the pandemic had a positive influence on pet owners. Pet ownership was associated with improved mood [ 72 ], reduced loneliness [ 73 , 74 ], greater social support [ 75 – 77 ], and relieved stress by increased physical exercise [ 75 , 78 ]. Dog owners reported that their dogs helped them cope with emotional stressors (91.2%) and maintain physical activity (96.4%) during lockdown [ 78 ]. Dog walking during confinements may have alleviated stressors and motivated self-care [ 73 ]. The results of a non-peer reviewed survey study that explored the relationship, experiences, and concerns of human-pet dog dyads during the pandemic suggested that dogs may help to reduce feelings of distress, anxiety, depression, and isolation [ 79 ]. However, recent studies have also reported that pet ownership during the COVID-19 pandemic may have negatively affected people because of limited availability to resources (e.g., veterinary care, pet supplies) [ 75 , 78 , 80 , 81 ].

The present study aimed to understand if pet dogs offered their owners social support and contributed to better wellbeing during the COVID-19 pandemic. It was hypothesized that pet dog ownership would act as a buffer against negative impacts caused by the pandemic. Two groups of people (pet dog owners and potential dog owners) were surveyed. Both groups were asked to answer validated psychometric questionnaires on depression, anxiety, and happiness. Other types of pets are also likely to provide social support to humans. However, it is unclear if this support is equivalent and if the psychological mechanisms involved are the same as human-dog relationships. In the context of the COVID-19 pandemic, there is emerging evidence that the relationship and attitude of people towards their pets may vary according to the species [ 82 ]. To avoid this potential confound, only dog owners and potential dog owners were included in this survey. Participants also answered questions about how much social support they perceived was available to them, and about their attitude towards and commitment to pets.

Participants

All participants were at least 18 years old, English speaking, had access to a computer, and were living in the United States at the time of the survey. Only one person per household was eligible to take the online survey. People owning other types of pets (e.g., cats, small rodents, rabbits, horses, birds), or who failed to complete the entire survey were excluded. Pet dog owners were defined as people who own at least one dog. Those who owned more than one dog were asked to answer the survey for the one dog that they felt the closest to. Because of the distinct nature of the relationship, we excluded dog owners whose dog was a service, emotional support, or therapy dog. Potential pet dog owners were defined as individuals who did not own a dog at the time of the survey but would be very or extremely interested in owning one in the future. People who did not have dogs and indicated that they were not interested in owning dogs in the future were excluded from the study.

Participants were asked to answer a series of demographic questions about race, ethnicity, employment status, education, annual household income, type of community they live in (urban, suburban, rural), number of dogs they own, age of the dog, and how long they have owned the dog (details can be found in S1 to S16 Tables). Participants in the two groups were then matched by sex, age, place of residence, and by the perceived negative impact that the pandemic had on their finances, emotions, health, and lifestyle.

A third-party research firm (Prodege, El Segundo, CA) recruited participants through multiple strategies including online advertisements via banner ads, referrals from current panelists, and targeted advertising in third-party search engines and websites across the Internet. The study was reviewed by an independent review board, Pearl Pathways (Indianapolis, IN), and received an exemption determination (protocol 20-NEST-101).

The survey was initially administered from November 9 th to the 24 th of 2020 (753 dog owners and 752 potential dog owners). Three hundred thirty-five (335) dog owners who also owned other pets were inadvertently allowed to complete the survey and were therefore removed from data analysis. In order to rebalance the groups, 335 potential dog owners with similar demographic profiles were also randomly removed from the data set. From February 18 th to the 22 nd of 2021, an additional 350 dog owners and 350 potential dog owners were surveyed. The final sample is composed of 1,535 participants (768 pet dog owners and 767 potential pet dog owners).

Psychometric scales

Participants were evaluated using the following six validated psychometric scales. The scales were selected based on their specialized assessments of human-pet relationships, perceived social support, and self-reported psychological states.

Pet Attitude Scale [ 83 ].

This 18-item scale measures favorableness of attitudes towards pets. Participants answer each item using a 7-point Likert scale (strongly disagree/strongly agree). The total score ranges from 18 (most unfavorable attitude towards pets) to 126 (most favorable attitude towards pets). This scale has high internal consistency and test-retest reliability [ 83 ].

Miller-Rada Commitment to Pets Scale [ 84 ].

This 10-item scale uses a 5-point Likert scale (strongly agree/strongly disagree) to assess how much time, energy, and resources an owner is willing to devote to their pet, even when facing a negative situation (e.g., pet’s destructiveness). The total score varies from 10 (least committed) to 50 points (most committed). The Miller-Rada Commitment to Pets Scale has high internal reliability [ 84 ].

Multidimensional Scale of Perceived Social Support (MSPSS) [ 85 ].

This 12-item, 7-point Likert scale (strongly agree/strongly disagree) instrument measures how much support people perceive they receive from friends, family, and significant others. To calculate a perceived social support score, the responses across all 12 items are summed and the total is divided by 12. There are no established population norms on the MSPSS. However, in this study, we adopted the rule proposed by [ 86 ]: scores ranging from 1 to 2.9 are considered low support; scores of 3 to 5 are considered moderate support; and scores greater than 5 are considered high support. The MSPSS is a psychometrically sound instrument with adequate internal and test-retest reliability, strong factorial validity, and moderate construct validity [ 85 ].

Center for Epidemiologic Studies Depression Scale-Revised (CESD-R) [ 87 ].

The 20 items of this 5-point Likert scale are based on the criteria for depression from the Diagnostic and Statistical Manual of Mental Disorders V [ 88 ]. Participants are asked how often they have experienced each item in the past two weeks (not at all or less than 1 day/nearly every day for 2 weeks). A score of less than 16 indicates little or no depression and is not considered clinically significant. The cut-off for possible depression is 16. The maximum score is 80. The CESD-R shows strong psychometric properties as demonstrated by its exploratory and confirmatory analyses, internal consistency, and convergent and divergent validity [ 89 ].

Generalized Anxiety Disorder Scale (GAD-7) [ 90 ].

This 7-item, 4-point Likert scale (not at all/nearly every day) measures the severity of generalized anxiety disorder in people. Participants are asked how often they have experienced each item in the past two weeks. Scores of 0 to 9 indicate no to mild anxiety. The cut-off score for anxiety is 10. The maximum score is 21. This scale is an efficient tool to screen for generalized anxiety disorder with high internal consistency, good test-retest reliability and procedural validity, strong diagnostic criterion validity and construct validity [ 90 ].

Oxford Happiness Questionnaire (OHQ) [ 91 ].

This scale provides a broad measure of personal happiness. It contains 29 items and participants answer each item using a 6-point Likert scale (strongly disagree/strongly agree). A personal happiness score, varying from 1 to 6, is obtained by summing the person’s answers and dividing the total by 29. The scores are interpreted as follows (adapted from [ 92 ]: less than 2 is “Not happy”; between 2 and 3 is “Somewhat unhappy”; between 3 and 4 is “Not particularly happy/unhappy”; between 4 and 5 is “Rather happy”; greater than 5 is “Very happy”. The OHQ has high internal consistency and good construct validity [ 93 ].

Sample size and power calculations

Based on published data from the CESD-R scale [ 89 ], the GAD-7 [ 94 , 95 ], and the OHQ [ 96 , 97 ], a sample size of at least 750 participants per group provides sufficient power (actual power 0.80) to detect a 15% difference between the groups (one-tail) at a confidence level of 95%.

Statistical analysis

Data transformation and descriptive statistics were computed using R [ 98 ]. The package openxlsx [ 99 ] was used to read and write excel files. The package rstatix [ 100 ] was used to perform Mann-Whitney U-tests. The data was composed of two non-normally distributed groups (confirmed with Shapiro test). Therefore, the Mann-Whitney U-test with Bonferroni correction was used (non-parametric alternative to an independent samples t-test) to assess the association between dog ownership and depression, anxiety, and happiness. It was also used to explore the possible links between dog ownership status, pet attachment and commitment, perceived social support, and depression, anxiety, and happiness. Effect size was calculated using Cohen’s d.

The descriptive statistics of the perceived impact of COVID-19 on finances, emotions, health, and lifestyle of the dog owners and potential dog owners are presented in S17 to S20 Tables. The descriptive statistics of the answers of the dog owners and potential dog owners to the six psychometric scales used in this study are presented in S21 to S26 Tables.

Perceived negative impact of the pandemic on the participants’ lives

Because the two groups of participants were matched based on their responses to perceived negative impacts of the pandemic, the overall results (dog owners and potential dog owners) are presented in this section. Thirty-three percent of the participants reported that their health had been somewhat to extremely impacted during this period. Forty-five percent indicated that their finances were somewhat to extremely impacted. Sixty-seven percent said that their emotions had been somewhat to extremely impacted. Finally, seventy-two percent reported that their lifestyle had been somewhat to extremely impacted.

Dog ownership status and depression, anxiety, and happiness

On average, dog owners had a lower depression score (M = 12.41; SD = 14.25), compared to potential dog owners (M = 14.06; SD = 14.86). The difference between the groups was statistically significant (U = 271493.5, p = .018) and the effect size was small (d = 0.07). There was no significant difference (U = 281667, p = .186) between the anxiety scores of the dog owners (M = 4.43; SD = 5.04) and the potential dog owners (M = 4.82; SD = 5.27). There was no significant difference (U = 305278, p = .216) between the happiness scores of the dog owners (M = 4.05; SD = 0.89) and the potential dog owners (M = 3.99; SD = 0.91).

Dog ownership status and attitude towards pets

Dog owners had a significantly more positive attitude towards pets on average compared to the potential dog owner group (U = 358322, p = .000). The average score of the dog owners on the Pet Attitude Scale was 109.93 (SD = 13.49) and the average score of the potential dog owners was 105.94 (SD = 13.16). The effect size of dog ownership on attitude scores was small (d = 0.19). There were no correlations between attitude towards pets and depression, anxiety, and happiness scores ( S27 Table ).

Dog ownership status and commitment towards pets

Dog owners, on average, showed significantly greater commitment to pets compared to potential dog owners (U = 355577.5, p = .000). The average score of the dog owners on the Miller-Rada Commitment to Pets Scale was 43.97 (SD = 7.25) and the average score of the potential dog owners was 41.49 (SD = 7.66). The effect size of dog ownership on commitment scores was small (d = 0.18). There were no correlations between commitment towards pets and depression, anxiety, and happiness scores ( S28 Table ).

Dog ownership and perceived social support

Dog owners reported having significantly more social support available to them compared to the potential dog owner group (U = 314088.5; p = .042). The average score of the dog owners on the Multidimensional Scale of Perceived Social Support was 5.48 (SD = 1.23) and it was 5.34 (SD = 1.28) for the potential dog owners. The effect size of dog ownership on perceived social support scores was small (d = 0.06).

Dog ownership, perceived social support, and depression, anxiety, and happiness

We explored ex post facto the possible links between dog ownership status, perceived social support, and depression, anxiety, and happiness by dividing dog owners and potential dog owners into groups according to their reported levels of social support: low, moderate, and high. There were no statistical differences in perceived social support and depression, anxiety, and happiness between dog owners and potential dog owners ( Table 1 ). However, the depression score’s mean value of the dog owners was lower than the depression score’s mean value of the potential dog owners for all three levels of social support. The anxiety scores for the dog owners who reported low and moderate levels of perceived social support were lower than the anxiety scores of the potential dog owners for the same levels of perceived social support. Further, the happiness scores of the dog owners were higher than the happiness scores of the potential dog owners for the low and moderate levels.

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

https://doi.org/10.1371/journal.pone.0260676.t001

This ex post facto exploration suggests that there was a correlation between levels of perceived social support and the depression, anxiety, and happiness scores of the dog owners and potential dog owners. It was not possible to analyze the data statistically because of the large differences between the number of participants in the reported levels of social support (77 in the low group, 420 in the moderate group, and 1032 in the high group). However, it is worth mentioning that the depression scores of the participants with low perceived social support were almost three times higher than the depression scores of the participants with high perceived social support (approximately 30 vs. 10). From a clinical perspective, it is also of significance to note that participants who reported low perceived social support had, on average, depression scores much higher than the cut-off for possible depression of 16 from the CESD-R [ 87 ]. Anxiety scores of the participants who reported low perceived social support were about 2.5 times higher than those who reported high perceived social support (approximately 9.6 vs. 3.7). All mean anxiety scores were lower than the cut-off of 10 of the GAD-7 [ 90 ]. However, the mean scores of the participants who reported low perceived social support were less than half a point from the cut-off value for possible anxiety. The OHS [ 91 ] mean happiness scores of the participants who reported low perceived social support were about 65% lower than those who reported high social support (approximately 2.8 vs. 4.3).

This study adds to the emerging body of literature on human-animal interaction during the COVID-19 pandemic by quantifying the potential benefits of pets on human psychological health and wellbeing. Our results showed that dog owners had significantly lower depression scores than potential dog owners, but the two groups had similar anxiety and happiness scores. In addition, dog owners showed significantly more positive attitude and higher commitment towards pets than potential dog owners. Dog owners also reported a significantly higher degree of perceived social support. Taken together, our results suggest that dog ownership may have provided people with a stronger sense of social support, which in turn may have helped buffer some of the negative psychological impacts caused by the COVID-19 pandemic.

In keeping with previously published research, our results showed the complexity of measuring the association between dog ownership and people’s wellbeing. While dog-ownership or -fostering has been correlated with better depression outcomes [ 101 , 102 ], other studies have found no correlations [ 54 , 58 , 64 , 103 , 104 ], or found that pet ownership was associated with higher depression levels [ 105 – 108 ]. In addition, the correlation between pet ownership and depression symptoms was modulated by marital status and sex [ 109 ]; unmarried men with pets had the highest depression scores amongst men, while married women with pets had the lowest depression scores amongst women.

Similarly, the correlation between pet ownership and owners’ anxiety symptoms is limited and inconclusive. Some exploratory studies have reported a link between pet ownership and decreased anxiety symptoms in children with autism [ 110 , 111 ] and in older adults [ 112 ]. However, other studies have reported no correlations [ 113 ], or increased anxiety in pet owners [ 106 ]. Investigating successful aging in older adults, one study reported that dog ownership was positively associated with psychological wellbeing indicators [ 114 ]. Still, most research on pet ownership and happiness and related concepts such as quality of life, wellbeing, or life satisfaction found no differences between pet owners and non-pet owners [ 58 , 106 , 115 – 118 ]. Our study found no link between dog ownership and anxiety and happiness scores. Multiple interfacing lifestyle factors such as physical activity, diet, sleep, stress management, avoidance of substance abuse, and social connection (including pet ownership) modulate depression, anxiety, and happiness. Teasing out their relative importance often proves to be difficult, and this may have been a factor at play in our study.

Our results showed that dog owners reported having more social support available to them compared to potential dog owners. These findings are consistent with other studies documenting that pets provide social support to humans [ 54 , 56 – 58 , 60 , 114 , 119 ]. However, when dog owners and potential dog owners were further divided according to their reported levels of social support, there was no difference between the groups in terms of depression, anxiety, and happiness. Also, while dog owners showed a more positive attitude towards and commitment to pets compared to the potential dog owners, the scores were not correlated to depression, anxiety, and happiness. In future research, it may be useful to look further into the quality of the relationship between people and their pets as it may be an important factor to consider. Different pet relationship scales may tap into different aspects of the human-pet relationship. For example, research that utilized Bowlby’s attachment theory [ 120 ] and Rogers’ core conditions [ 121 ] to examine human-pet relationships found a correlation between secure pet attachment and better quality of life, lower psychological distress, and better psychopathology in pet owners [ 118 ]. In a study conducted in the first half of 2020 with cat and dog owners, above average pet attachment was found to be a protective mental health factor for people with moderate to high levels of mental health symptoms. However, high pet attachment in people with severe levels of mental health symptoms was associated with decreased chances of improved mental health [ 122 ]. The pet relationship scales used in the current study may not have facilitated insight into the nature of the human-pet relationship that most affects depression, anxiety, and happiness during a crisis such as the COVID-19 pandemic.

The dog owner participants were not asked questions regarding the type of activities they engaged in with their dogs. Doing so may have helped better measure the benefits of owning a dog. In response to the prevalence of contradictory findings in pet ownership benefit studies, a framework was developed to isolate variables within the pet ownership lifestyle by focusing on specific activities associated with pet ownership that could promote clearer results [ 123 ]. Other findings suggested that future research could also investigate how specific activities (e.g., dog walking, tactile interaction, meeting pet’s needs) modulate mental health measures like depression, anxiety, and happiness among pet owners [ 56 ].

There was a difference between some of our results and what pet owners say when asked about the importance of their pet during the COVID-19 pandemic. For example, in a recent bespoke questionnaire, 58% of respondents reported that their dog helped manage anxiety and 57% reported their dog helped with depression [ 124 ]. Seventy-two percent of pet owners indicated that they “would not have been able to get through” the pandemic without their pet [ 125 ]. Eighty-seven percent of people indicated that their animal was helping them to cope emotionally during the pandemic; 91% reported that their pets were a significant source of emotional support; and 94% said that their animal had a positive effect on their family [ 78 ]. Based on these results, it appears that pets were highly valued and that they positively contributed to the quality of life of their owners in meaningful ways during the pandemic. Why was the dog effect not more evident in the data we collected? One possible explanation is that its effect is real, but its effect size is smaller than the effect of other lifestyle factors that may affect psychological outcomes. Also, when prompted about the role that their dogs may play in specific situations, people may overestimate their actual impact [ 126 ]. In all likelihood, people are genuine when they report how much they love their dogs and how much comfort their dogs bring them. However, the dog effect may not be strong enough to completely counterbalance the traumatic impact of major life events such as the loss of a job, a divorce, or, in our case, the COVID-19 pandemic. Similarly, no quantitative correlations were found between pet ownership and loneliness and wellbeing, despite people qualitatively commenting that their pets provided them with social support and a sense of purpose during the pandemic [ 82 ].

Pets’ contribution to the wellbeing of people may be more apparent among individuals in precarious states (e.g., high stress, socially isolated) [ 52 ]. In our study, 70% of dog owners (and 65% of potential dog owners) reported benefiting from high social support from family and friends and this degree of social support is likely to have provided a buffering effect against the negative impact of the COVID-19 pandemic. Dogs likely contributed positively to the general wellbeing of the dog owner participants, but this effect may have been masked by the social support received by the participants from other people. Therefore, future research focusing on people with low and moderate social support may better capture differences between dog owners and potential dog owners.

An association between levels of perceived social support and average depression, anxiety, and happiness scores among all participants (both dog owners and potential dog owners) was observed in an ex post facto analysis that we conducted. Compared to those who reported high perceived social support, people who reported low social support had depression and anxiety scores about twice as high and their happiness scores were notably lower. It is important to point out that this association is solely based on the visual comparison of the mean scores. Yet, this observation is in line with the body of literature on the importance of social support in difficult times [ 27 , 31 , 32 , 34 ]. These observations support the suggestion that dog ownership effects might be most measurable within populations of people with low to moderate social support.

The current study was carefully designed to avoid common methodological mistakes in human-animal interaction research [ 127 ], but still, some limitations are present. The timing of survey studies is always an important factor to consider. The survey was initially administered in November 2020, a month of anticipation and stress for many Americans headed into the holiday season. It was administered a second time in the spring of 2021. While there were concerns about the effect of the timing differences between the two survey groups, carefully controlled recruitment indicates that there were no demographic differences between the groups. Of note, the US government only issued voluntary public health recommendations aimed at controlling the transmission of the virus. There were no federally unified or enforced responses. What the implemented mandates were and when they were implemented varied by states and by localities [ 128 ]. It is possible to imagine that people who lived in states or localities where the restrictions were more extensive or lasted longer would have benefited differently from the presence of their dogs compared to people who lived in places where the restrictions were fewer or shorter in duration. Because of this great variability, we were not able to analyze potential differences between areas.

A large number of participants were used, and potential confounders were controlled for (e.g., sex, age, place of residence, perceived negative impact that the pandemic had on finances, emotions, health, and lifestyle). Participants in both groups were also balanced on all demographic variables. The limitation of participants’ self-reported information is acknowledged, however validated psychometric tools that have been widely and successfully used by many researchers were used. To avoid selection bias, participants were recruited by an independent third-party.

It was not possible to use a true experimental design where half the participants would be provided with a dog and half would not get a dog. Usually, research investigating the possible benefits of dog ownership compare a group of dog owners to a group of non-dog owners. Instead, a quasi-experimental design was developed such that dog owners were compared to a specific group of non-dog owners, i.e., people who did not own a dog at the time of the survey but who indicated that they were very interested in acquiring one. This selection criteria may offer a more effective comparison than including a group of participants who are not interested in dogs and who made the decision not to live with one. It is fair to assume that these people would gain no benefit from owning pet dogs and that they seek and receive social support through other means. In future research, we propose that in addition to comparing dog owners to potential dog owners, consideration should be given to including participants with more diverse pet attachment and commitment scores. In this sample population, most participants scored high on both measures and this may not be representative of the general population.

The COVID-19 pandemic has negatively affected diverse populations [ 129 – 134 ] and our results provide evidence that pet owners and potential pet owners have also been impacted. Our results show that pet dog owners were significantly less depressed than non-pet owners during the COVID-19 pandemic. They are attached and committed to their dogs and they reported more social support available to them. Our work adds to the corpus of scientific literature demonstrating that pet dogs may positively contribute to the wellbeing of owners during difficult times. However, more work is needed to better understand the relationship between pet ownership and social support as modulators of owner wellbeing. Future research should focus on people with low and moderate social support and include owners with diverse dog attachment levels.

Supporting information

S1 table. age of participants..

https://doi.org/10.1371/journal.pone.0260676.s001

S2 Table. Sex/gender of participants.

https://doi.org/10.1371/journal.pone.0260676.s002

S3 Table. US state of residence.

https://doi.org/10.1371/journal.pone.0260676.s003

S4 Table. US division of residence.

https://doi.org/10.1371/journal.pone.0260676.s004

S5 Table. US region of residence.

https://doi.org/10.1371/journal.pone.0260676.s005

S6 Table. Number of dogs.

https://doi.org/10.1371/journal.pone.0260676.s006

S7 Table. Ownership duration.

https://doi.org/10.1371/journal.pone.0260676.s007

S8 Table. Age of dogs.

https://doi.org/10.1371/journal.pone.0260676.s008

S9 Table. Number of people in household.

https://doi.org/10.1371/journal.pone.0260676.s009

S10 Table. Marital status.

https://doi.org/10.1371/journal.pone.0260676.s010

S11 Table. Education level.

https://doi.org/10.1371/journal.pone.0260676.s011

S12 Table. Employment status.

https://doi.org/10.1371/journal.pone.0260676.s012

S13 Table. Annual income.

https://doi.org/10.1371/journal.pone.0260676.s013

S14 Table. Metropolitan-nonmetropolitan classification.

https://doi.org/10.1371/journal.pone.0260676.s014

S15 Table. Racial or ethnic background.

https://doi.org/10.1371/journal.pone.0260676.s015

S16 Table. Hispanic origin or descent.

https://doi.org/10.1371/journal.pone.0260676.s016

S17 Table. Perceived impact of Covid-19 on finances.

https://doi.org/10.1371/journal.pone.0260676.s017

S18 Table. Perceived impact of Covid-19 on emotions.

https://doi.org/10.1371/journal.pone.0260676.s018

S19 Table. Perceived impact of Covid-19 on health.

https://doi.org/10.1371/journal.pone.0260676.s019

S20 Table. Perceived impact of Covid-19 on lifestyle.

https://doi.org/10.1371/journal.pone.0260676.s020

S21 Table. Pet Attitude Scale descriptive statistics.

https://doi.org/10.1371/journal.pone.0260676.s021

S22 Table. Miller-Rada Commitment to Pets Scale descriptive statistics.

https://doi.org/10.1371/journal.pone.0260676.s022

S23 Table. Multidimensional Scale of Perceived Social Support descriptive statistics.

https://doi.org/10.1371/journal.pone.0260676.s023

S24 Table. Center for Epidemiologic Studies Depression Scale-Revised descriptive statistics.

https://doi.org/10.1371/journal.pone.0260676.s024

S25 Table. Generalized Anxiety Disorder Scale descriptive statistics.

https://doi.org/10.1371/journal.pone.0260676.s025

S26 Table. Oxford Happiness Scale descriptive statistics.

https://doi.org/10.1371/journal.pone.0260676.s026

S27 Table. Correlations between attitude towards pets and depression, anxiety, and happiness scores.

https://doi.org/10.1371/journal.pone.0260676.s027

S28 Table. Correlations between commitment to pets and depression, anxiety, and happiness scores.

https://doi.org/10.1371/journal.pone.0260676.s028

https://doi.org/10.1371/journal.pone.0260676.s029

• View Article
• PubMed/NCBI
• Google Scholar
• 21. Wills TA. Social support and interpersonal relationships. Prosocial behavior. Review of personality and social psychology, Vol. 12. Thousand Oaks, CA, US: Sage Publications, Inc; 1991. p. 265–89.
• 37. McNicholas J, Collis GM. The End of a Relationship: Coping with Pet Loss. In: Robinson I, editor. The Waltham Book of Human–Animal Interaction: Elsevier; 1995. p. 127–43.
• 41. Bonas S MJ, Collis G. Pets in the network of family relationships: An empirical study. In: Podberscek AL PE, Serpell JA, editor. Companion animals and us: Exploring the relationships between people and pets Cambridge: Cambridge University Press; 2000. p. 209–36.
• 60. Pachana NA MB, Robleda-Gomez S. A developmental psychological perspective on the human-animal bond. Blazina C, Boyraz G., Shen-Miller D., editor. New York: Springer; 2011.
• 79. Fido of Fort Collins. Pet dogs during the time of COVID [Available from: https://fidofortcollins.org/covid-dog-owner-survey-results/ ].
• 80. Kim K. Anxious pet owners face delays getting veterinarian appointments, even for sick animals. The Washington Post. 2020 12/08/2020.
• 86. Research CoEfHD. Multidimensional Scale of Perceived Social Support (MSPSS) [Available from: https://elcentro.sonhs.miami.edu/research/measures-library/mspss/index.html .
• 88. Association AP. Diagnostic and statistical manual of mental disorders. 5th ed. Washington D.C.: American Psychiatric Association; 2013.
• 98. R Core Team R. R: A language and environment for statistical computing. R foundation for statistical computing Vienna, Austria; 2013.
• 120. Bowlby J. Attachment and loss. New York: Basic Books; 1969. https://doi.org/10.1001/archpsyc.1969.01740180074007 pmid:4890814
• 121. Rogers CR. Client-centered therapy: Its current practice, implications and theory. Boston: Houghton Mifflin; 1951.
• 125. SWNS Staff. Study finds that having pets around is helping most Americans get through the pandemic. https://swnsdigital.com/2020/08/study-finds-that-having-pets-around-is-helping-most-americans-get-through-the-pandemic/ ). [Accessed 2021 Nov 08].

Dogs Supporting Human Health and Well-Being: A Biopsychosocial Approach

Affiliations.

• 1 Department of Psychiatry, Center for Human Animal Interaction, School of Medicine, Virginia Commonwealth University, Richmond, VA, United States.
• 2 Human-Animal Bond in Colorado, School of Social Work, Colorado State University, Fort Collins, CO, United States.
• 3 Department of Education, California State Polytechnic University, Pomona, CA, United States.
• 4 Division of Social Sciences and Natural Sciences, Seaver College, Pepperdine University, Malibu, CA, United States.
• PMID: 33860004
• PMCID: PMC8042315
• DOI: 10.3389/fvets.2021.630465

Humans have long realized that dogs can be helpful, in a number of ways, to achieving important goals. This is evident from our earliest interactions involving the shared goal of avoiding predators and acquiring food, to our more recent inclusion of dogs in a variety of contexts including therapeutic and educational settings. This paper utilizes a longstanding theoretical framework- the biopsychosocial model- to contextualize the existing research on a broad spectrum of settings and populations in which dogs have been included as an adjunct or complementary therapy to improve some aspect of human health and well-being. A wide variety of evidence is considered within key topical areas including cognition, learning disorders, neurotypical and neurodiverse populations, mental and physical health, and disabilities. A dynamic version of the biopsychosocial model is used to organize and discuss the findings, to consider how possible mechanisms of action may impact overall human health and well-being, and to frame and guide future research questions and investigations.

Keywords: biopsychosocial; canine; dog; human health; human-animal interaction; mental health.

Copyright © 2021 Gee, Rodriguez, Fine and Trammell.

Publication types

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 26 June 2020

Mental representation and episodic-like memory of own actions in dogs

• Claudia Fugazza 1 ,
• Péter Pongrácz   ORCID: orcid.org/0000-0001-5126-299X 1 ,
• Ákos Pogány   ORCID: orcid.org/0000-0001-9498-0158 1 ,
• Rita Lenkei 1 &
• Ádám Miklósi 1 , 2  

Scientific Reports volume  10 , Article number:  10449 ( 2020 ) Cite this article

9824 Accesses

16 Citations

157 Altmetric

Metrics details

We investigated whether dogs remember their spontaneous past actions relying on episodic-like memory. Dogs were trained to repeat a small set of actions upon request. Then we tested them on their ability to repeat other actions produced by themselves, including actions performed spontaneously in everyday situations. Dogs repeated their own actions after delays ranging from a few seconds to 1 hour, with their performance showing a decay typical of episodic memory. The combined evidence of representing own actions and using episodic-like memory to recall them suggests a far more complex representation of a key feature of the self than previously attributed to dogs. Our method is applicable to various species, paving the way for comparative investigations on the evolution and complexity of self-representation.

Similar content being viewed by others

Memory for own actions in parrots

Sara Torres Ortiz, Simeon Q. Smeele, … Auguste M. P. von Bayern

Contextual inference underlies the learning of sensorimotor repertoires

James B. Heald, Máté Lengyel & Daniel M. Wolpert

Rapid learning of object names in dogs

Claudia Fugazza, Attila Andics, … Ádám Miklósi

Introduction

The cognitive ability to mentally represent own actions in non-human species is a challenging topic, elusive from the methodological point of view, mainly due to the difficulty of devising methods to test it under ecologically relevant conditions. Dolphins demonstrated the ability to repeat their own actions 1 and to perform an action that differed from the immediately preceding one 2 . However, as a result of their previous training, they expected the test and therefore, their performance may have relied on a prepared behavioural response. Consequently, it remains unknown whether they formed a specific mental representation of their own past actions.

The ability to mentally represent own actions may constitute one of the main building blocks of the more complex ability to represent the self 1 , 3 . In this view, self-representation is composed by an array of cognitive traits that may have evolved unequally in different species as a result of different selection processes during their evolution for a similar argument see 4 . Self representation has been widely studied in humans and in other primates e.g. 5 , 6 , 7 , but whether non-human animals represent any aspect of the self remains a controversial topic. Recent findings on metacognition in non-primate species e.g. rats 8 ; dolphins 9 challenge the view that any form of self-representation evolved only in primates. The traditional methodological approach to study self-representation is the mirror self-recognition paradigm that has been applied not only to apes 7 , but also to other species (including elephants 10 ; dolphins 11 ; magpies 12 cleaner wrasses 13 ). However, this paradigm may reveal only a single aspect of self-representation: visual recognition of one’s own image (cf. mirror-guided body inspection 14 ). Moreover, the view of self-representation as an undividable cognitive trait does not disclose its complexity 15 , 16 , as self-representation extends beyond the visual domain 17 . We argue that, based on evolutionary history and ecology, animal species should differ in the complexity of the mental representation that they may construct about themselves.

Dogs, owing to the social richness of their environment, offer an ideal model species to address these questions. Although dogs can use a mirror to solve a task e.g. 18 , evidence of mirror self-recognition in this species has not been provided and, to date, there is no conclusive evidence about any aspect of self-representation in dogs. We argue that for canines, recognizing own appearance (e.g. color of the fur or presence of a spot on it) may not be of any particular advantage, thus it is unlikely that they evolved representation of their own physical appearence. We suggest that there is a need for a wider array of test paradigms for testing different building blocks of self-representation, and the ecology of the species should be taken into account when selecting the applied method, instead of the current anthropocentric and simplistic approach 19 . Some of the previous studies on dogs made steps along this line of thought. For instance, recently it was found that dogs approach an opening suitable for their body size faster than a too small one, suggesting that they represent their body size 19 . A different study applied a modified version of the mirror-self recognition test focusing on olfaction to test dogs’ ability to recognize their own urine 20 and showed increased investigation of own modified scent. However, the results should not be considered as conclusive evidence of self-recognition because alternative explanations have not been excluded (e.g. increased investigation may be a ‘surprise’ reaction to a novel feature of a stimulus to which dogs are completely habituated).

Some building blocks of self represention might have contributed to dogs’ success in living in human social groups cf. 21 . Particularly, representation and memorization of own actions may be among the most relevant features of the self for dogs living in the complex and variable social environment provided by humans.

Recently, we provided experimental evidence of episodic-like memory in dogs 22 . Episodic memory is a memory of personal events and episodes in one’s life and, in humans, it is thought to be linked to self-representation because it implies the ability to represent the self in the past e.g. 17 , 23 , 24 . This, in turn, may also allow an individual to plan future actions 25 . The existence of episodic-like memory in non-human animals has been investigated using different methodologies that reflect diverse theoretical approaches to its definition. While some authors define episodic memory from the perspective of the content of what is remembered, identifying episodic memory as memory of the what , where and when of an event 26 , other authors define episodic memory based on the mechanism how the past event is encoded and recalled. Accordingly, recall of an event relies on episodic memory when such event was encoded incidentally, that is, without knowing that it had to be remembered 27 , 28 . Episodic-like memory has been investigated in mammals and birds with both approaches. The research on episodic memory in non-human species was initiated by investigating memory of what, where and when in birds by Clayton and Dickinson 1998 29 who studied cache-recovery behaviour in scrub jays. Evidence of memory of the object and spatial component of episodic memory ( what and where ) has also been documented in one dog (Kaminski 2008) 30 . From the perspective of the type of encoding and recall, episodic memory, defined as memory of events encoded incidentally, has been found in chimpanzees and orangutans 31 , rats 27 , 32 , pigeons 28 and dogs 22 .

A crucial methodological criterion of experiments on memory of events that were encoded incidentally is the unexpectedness of the recall test 33 . This ensures that there has been no reason for explicit encoding (i.e. enconding information that is known to be important), hence incidental encoding can be reasonably assumed. As the mental experience of memory retrieval is subjective and, therefore, it is not possible to directly assess it in non-human species, along with other authors e.g. 28 , 32 , we refer to memory of incidentally encoded events as episodic-like memory.

Here we aimed at testing mental representation and episodic-like memory of own spontaneous actions in dogs. We refer to mental representation as the ability to represent things that are not present to the senses and to episodic-like memory as memory of events (i.e. own actions in our study) enconded incidentally 27 , 28 . Preliminarily, we tested the ability of dogs trained to repeat their own actions to generalize this rule and repeat also actions not included in the training. These tests resembled the training situation in which, typically, the request to perfom an action was followed by the repeat command. Therefore, after receiving a command to perfom an action, the dogs could expect to be asked to repeat and, consequently, encoding of the perfomed action might have been explicit and the dogs’ performance could have relied on a prepared behavioural response.

As our primary aim was to test whether dogs represent actions performed by themsleves spontaneously in everyday-life contexts (i.e., out of a trainig context) and recall those relying on episodic-like memory, we also tested them in situations when the repeat test was unexpected. These tests were carried out in everyday-life situations that did not resemble training and, importantly, the dogs were not given any command to perfom an action but, instead, their owners apparently ignored them. The (unexpected) repeat command was given as soon as the dog spontaneously perfomed a well-identifiable action (e.g. drank water from its bowl or lied down).

In our previous study on episodic-like memory 22 the tests were carried out in a training/testing context and the test was preceded by a human demonstration of an action, necessarily. Thus, to prevent dogs from encoding the episodes explicitly, we had to ‘re-train’ them to modify their expectation of the test. It has been argued 34 that, although evidence of the unexpectedness of the memory assessment was compelling, we cannot be certain that encoding was incidental because, in the training history of the dogs, action demonstrations were reliably followed by an imitation test. Consequently, action demonstrations may have been explicitly encoded, in anticipation of the memory assessment. To address this issue, in the key tests of the current study no previous commands are given to the dogs and the situation does not resemble a training context. This ensures that the test was completely unexpected and encoding of the actions performed must have been undoubtedly incidental 33 , 35 .

Potentially expected tests

We applied a specific training procedure (Repeat training, see supporting information), to teach 10 dogs to reproduce their own actions on command ‘Repeat!’. We assessed the success of the repeat training in a Baseline repeat test during which the owner asked his/her dog to perform and then repeat the actions used for the Repeat training. This test was done to verify the success of the training process and to assess the baseline level of success of dogs repeating trained own actions. The dogs (N = 10) successfully repeated their own actions irrespectively of whether those were familiar ( Baseline ) or this was the first time they were asked to repeat those ( Untrained actions repeat ), when they could potentially have expected the repeat test (Likelihood Ratio Test (LRT), effect of Experimental condition: χ 2 1  = 2.37, p = 0.124; Table  1 ). These two tests confirmed, therefore, that our Repeat training (see supporting information) was successful.

To test how flexibly the dogs could generalize the repeat rule, dogs were asked to repeat staying still in a sitting position in the Doing nothing test. In this test, 9 (of 10) dogs stayed in their position for at least 5 s (thus repeated ‘staying’ or ‘doing nothing’). The remaining one dog lied down 4 s after the repeat command was given.

We controlled for potential unvoluntary cues driving the preformance of the dogs in the Clever Hans control test a test. All 10 dogs repeated their own actions when the command was given by person who did not know the past action of the dog, excluding this explanation for the dogs’ performance.

In the ‘ Who is acting’ test the dog was randomly asked to either repeat its own action, or to imitate the action demonstrated by its owner or first imitate an action demonstrated by its owner and then to repeat its own action. The dogs’ performance (88.3% successful trials) did not differ from the Baseline, irrespective of whether the owners issued the repeat command, the Do it command or the repeat command following the ‘Do it’ command - LRT of Experimental condition: χ 2 1  = 0.88, p = 0.348 (Fig.  1 ).

Proportion of successfully repeated own actions or actions demonstrated by the owner in the Baseline repeat test (open circle) and in the ‘ Who is acting’ test (filled symbols). The graph shows responses of the 10 dogs in different test conditions: in the Do it trials (Who-D; the owner demonstrated an action and gave the ‘do it’ command); in the repeat trials (Who-R; the owner asked the dog to perform a trained action and then gave the repeat command); and in the Do it + repeat trials (Who-DR; the owner demonstrated an action, gave the ‘do it’ command and, after the dog performed an action - i.e. imitated - gave the repeat command).

Unexpected tests

In the unexpected tests, dogs were able to repeat actions that they performed spontaneously, in everyday situations ( Spontaneous action test and Spontaneous object-action test) . In the Spontaneous action test , the repeat command was issued by the owner while sitting on a bench or sofa (i.e. while apparently ignoring the dog that was free in the area) as soon as the dog spontaneously performed a well-identifiable action (e.g., jumped on the sofa, sat, laid on the floor, drank water from its bowl etc.). In the Spontaneous object-action test , novel objects were provided that the dog was free to explore (i.e., the objects were on the floor) and the repeat command was issued after an identifiable action with any of the objects was observed. We did not find significant differences between the subjects’ ability to repeat own actions in any of the Spontaneous tests, compared to the Baseline (LRT of Experimental condition, Spontaneous action vs Baseline : χ 2 1  = 1.14, p = 0.286; Spontaneous object-action vs Baseline : χ 2 1  = 0.11, p = 0.738; Table  1 , Fig.  2 ). In the Spontaneous action test and Spontaneous object-action test , 7 of 10 dogs repeated their own actions.

Proportion of successfully repeated actions of 10 dogs in the different experimental conditions, including control tests. Open circles represent the Baseline condition (BL-R), in which the dog was asked to repeat actions that were used during the Repeat training. On panel a), success rates during the Spontaneous action test (SA-R, black filled circles) and two corresponding control tests: the Spontaneous action control (when no command was issued but the dog was let free again in the test area; SA-C, black diamond) and the Different word control (when instead of repeat, a different word was said; SA-W, black triangle) are illustrated. On panel b), success rates during the Spontaneous object-action test (SOA-R, black filled circles) and the Spontaneous action control (when no command was issued; SOA-C, black diamond) are illustrated. In the two Spontaneous repeat tests (SA-R and SOA-R) dogs were tested with different delays (0 sec, 20 sec, 1 min, 1 h) between their spontaneous action and requesting to recall and repeat it.

We ensured that dogs would not simply perform a situation-specific action in a given situation (i.e., even if not requested to repeat), by reintroducing the dog in the same context whithout giving the repeat command. In the Spontaneous action control and Spontaneous object-action control tests none of the dogs repeated the action they had performed previously, thereby excluding that they would always perform those actions in that context. Dogs showed various actions during this test, including sniffing on the floor, licking the owner’s face, lying down and interacting with different objects in the Spontaneous object-action control .

In line with the above controls, in the Different word control trials, when a different word of no meaning to the dogs was uttered instead of the repeat command, none of the dogs repeated the previously performed action. All dogs except for two, each in one trial, stood in their position for at least 5 s before starting other activities after hearing the meaningless word. One dog lied down and another dog looked at the owner and walked away.

We also tested the dogs by including delays of different lengths (20 s, 1 min, 1 h) between the actions produced spontaneously and the command to repeat in the spontaneous tests. During the delay the dog was taken away from the context where it performed the action and was kept in a crate, thereby limiting the other behaviours that it may spontaneously perform. Dogs’ success decreased gradually with time. With 20 s delay, dogs’ performance was not significantly different from Baseline in either test conditions (LRT of effect of delays, 20 s vs Baseline , Spontaneous action test and Spontaneous action-object test , both p > 0.286). When the delay was 1 min, dogs’ performance decreased further (LRT of effect of delays, 1 min delay vs Baseline , Spontaneous action test : χ 2 1  = 3.026, p = 0.082; Spontaneous object-action test : χ 2 1  = 5.645, p = 0.018). In both test conditions, when tested after the longer delay of 1 h, success to recall own actions was lower than in the Baseline (LRT of effect of delays, 1 h delay vs Baseline , Spontaneous action test: χ 2 1  = 13.03, p < 0.001; Spontaneous object-action test : χ 2 1  = 8.97, p = 0.003; Fig.  1 ). Specifically, in the Spontaneous action test with 20 s delay, 7 of 10 dogs repeated their own actions; with 1 min delay 6 dogs, whereas with 1 h delay 3 of 10 dogs were successful. In the Spontaneous object action test with 20 s delay 7 of 10 dogs repeated their own actions; with 1 min delay 5, and with 1 h delay 4 of 10 dogs were successful.

In the Control for context test the repeat command was given in a different context, in which none of the dogs (N = 4) repeated their own actions.

We provide experimental evidence of mental representation of own actions and episodic-like memory to recall them, as assessed by the means of unexpected recall tests, revealing dogs’ ability to represent spontaneous actions performed by themselves in the past. Dogs were able to repeat their own actions, irrespectively of whether these were included in the ‘repeat training’ or not. Transfer tests of this kind, in which successful performance on one cognitive task is applied to another, ensure that the subjects learned a rule and not a stimulus-response association 36 . Moreover, dogs were also able to repeat their own actions in tests where the command to repeat was alternated with the command to imitate others’ actions and in trials in which they were asked to ‘stay’ (i.e., ‘do nothing’), thereby showing flexibility in generalizing the ‘repeat rule’ to different situations.

Most importantly, dogs were able to repeat their own actions when the command to repeat was issued unexpectedly, in everyday-life contexts. When trained animals are tested in a test that resembles the training situation, it is difficult to determine whether mental encoding is incidental and the test is indeed unexpected 22 . By testing dogs in everyday-life situations that did not resemble training or testing sitations, we ensured that encoding of their own actions was incidental and, consequently, that remembering those must have involved the using of episodic-like memory. We can, thereby, exclude the possibility that the dogs’ performance relied on a prepared behavioural response, in contrast with previous studies in non-human species that did not rule out this possibility e.g. dolphins 1 , 2 .

The numerous controls that we ran in this study convincingly exclude that the dogs would always perform those actions in those situations and make it less likely that the prior experience of having performed a given action in that context simply increased the probability of that response due to priming 37 .

The dogs remembered and repeated their own spontaneous actions also after delays of 20 seconds and, in the case of the Spontaneous action test , after 1 minute. Some dogs recalled their own actions even after 1 hour. Success in these tests shows that the dogs had formed a mental representation of their own previous action and used it as the basis for performing the same action again, after some delay, when unexpectedly asked to do so. We argue that the combined evidence of mental representation of own spontaneous actions and episodic-like memory to recall them, as assessed by the means of unexpected recall tests, reveals a far more complex representation of a key feature of the self than previously attributed to dogs.

As expected 38 , dogs’ memory of their own actions presented a steep deterioration with increasing delays, showing a decay that is in line with previous studies in which recall recruited episodic memory 22 , further supporting that recall in these tests relied on this type of memory (cf. 22 and 39 , in which expected/unexpected memory tests were carried out with delays similar to those in the present study). In the test where the context was changed between encoding and recall, dogs were not successful in repeating their own actions. Decreased performance when context is changed between encoding and retrieval has also been found in the case of human episodic memory (e.g. 40 , 41 . However, for this test we only had the opportunity to test 4 subjects, therefore a relatively small sample size warrants caution when interpreting these findings.

We suggest that the ability to represent and remember own actions relying on episodic memory is an important building block of the more complex ability to represent ‘the self’. Such building block of self-representation may have evolved earlier in evolution and could be more widespread than the more complex conceptual knowledge about the self 42 . Its presence, however, does not imply also the presence of autonoetic consciousness see also 43 . Thus, we do not argue that these results show a fully fledged conceptual knowledge of the self in dogs. We rather argue that representation of own actions, one form of representation of the self, has evolved in this species, that is phylogenetically distant from humans, and may have evolved in a range of other social animals too.

We acknowledge that this study was conducted on a relatively small number of dogs; both the extensive efforts of training the subjects with the two methods applied and testing them in numerous control conditions to exclude alternative explanations limited further subject recruitment. Nevertheless, we argue that our results provide solid evidence to support our conclusions. Our methodological approach has the potential to be applied to other species. This paves the way to new research on the evolution of functionally equivalent abilities, by shedding light on mental self-representation in humans as an array of cognitive traits that may reveal a mosaic evolutionary pattern in different species 44 . These features of the self-representation eventually might have converged and interlocked in humans into a versatile ability to be aware of the self.

We applied a specific training procedure (Repeat training, see supporting information), to teach 10 dogs to reproduce their own actions on command ‘Repeat!’. The dogs were also trained to imitate actions demonstrated by a human with the Do as I Do method 45 (Supporting information). Dogs were tested in the following 12 tests, always starting with the Baseline test followed by the other tests in random order, apart from the Control for context test (in which only four dogs participated), that was the last test administered. Apart from this last test, each dog participated in every test.

Baseline repeat test

The Baseline repeat test consisted of 12 trials during which the owner asked his/her dog to perform and then repeat the six actions used for the Repeat training. At the beginning of each trial, the owner placed his/her dog in front of him/herself, then requested the dog to perform a predetermined action using commands known by the dog. As soon as the dog completed the command, the owner led the dog back to the starting position in front of him/herself and gave the repeat command while looking straight ahead, thus avoiding to give any inadvertent cues. After the dog had performed an action, irrespectively of whether it correctly repeated the previous one or performed something different, the test continued with the next trial, once the owner had repositioned the dog again in the starting position in front of him/her. The order of the six requested actions was semi-randomized so that every action was requested two times within the 12 trials.

Untrained actions repeat test

The dog was asked to perform actions it was already trained to perform, but had never been asked to repeat. The procedure was identical to that of the Baseline repeat test , but the requested actions differed from those included in the Repeat training. As the already-trained actions varied from dog to dog due to their training history, the type and number of actions used in the test also varied. The number of actions in which the dogs were tested varied from 4 to 8 (mean ± SD = 6.3 ± 0.6) Example of actions included: sit, touch a cone, bark, enter the agility tunnel, jump over a hurdle, spin.

Doing nothing test

The owner asked the dog to stay in a sitting position (‘Stay!’, i.e. do not move) using cues known by the dogs and waited for 5 s. Then the owner gave the repeat command. The behaviour of the dog was recorded for the following 20 s. This test consisted of one trial per dog.

Clever Hans control

The owner and another person, familiar to the dog, stood back to back. The familiar person attracted the dogs’ attention and called the dog in front of him/her. From this moment on, the owner closed his/her eyes and ears and sang a song loudly in order to ensure that s/he would not hear any noise potentially made by the dog. The familiar person asked the dog to perform a trained action using only visual cues (i.e. gestures). Immediately after, the familiar person touched the owner’s side with hand as an agreed signal to exchange position so that the owner now moved in front of the dog, that typically had returned to its original position after performing the requested action. The owner opened his/her eyes and ears and gave the repeat command without knowing what action the dog had previously performed. This test consisted of one trial per dog.

‘Who is acting’ test

In this test each dog was asked to either repeat its own actions (Repeat trials: R) or to imitate actions demonstrated by its owner (Do it trials: D) in a single 12-trial session. The test procedure in the R-trials was the same as described above for the baseline test. In the D-trials, the owner demonstrated an action, then asked the dog to imitate it (‘Do it!’). We also included trials in which the owner demonstrated an action, asked the dog to imitate and then asked the dog to repeat (Do it + Repeat trials: DR). Note that every DR trial included two actions by the dog: the imitation of the action demonstrated by the owner and then the repetition of the action just performed by itself. The order of the type of trials was semi-randomized and was the same for each dog: R, DR, R, D, R, D, DR, R, DR. The actions demonstrated or requested were those included in the baseline test.

We tested the dogs’ ability to reproduce their own spontaneous actions in everyday-life situations that did not resemble training/testing situations, in absence of a command to perform a specific action.

Spontaneous action test

The owner sat on a bench (or chair or sofa) in a place that was familiar to the dog: its house or outdoor area, based on areas suitable for the test that were available for the dog owners and familiar for the dogs. The owner was instructed to type on his/her mobile phone or to read a book, thus apparently ignored the dog that had the opportunity to move freely in the area. As soon as the dog spontaneously preformed a well-identifiable action (e.g., lied down, drank water from a bowl, jumped on a sofa), the owner called the dog in front of him/her and gave the repeat command. The repeat command was always given by the owner while looking straight ahead, in order to avoid inadvertent cues (see also Clever Hans control above ). To prevent dogs from forming expectations about being tested due to repeated exposure to the tests, we carried out one such trial per dog. These situations were not practiced during training.

Spontaneous object-action test

Four objects that were novel for the dog – a wooden statue in shape of an animal, a plush toy, a dog crate and a doll – were placed in a familiar area (house or outdoor area) at 50 cm from each other.

The owner approached the area with the dog unleashed and let the dog to explore freely, while the owner remained passive. As soon as the dog spontaneously preformed a well-identifiable action (e.g., touched an object with its paw, grabbed an object with its mouth), the owner called the dog back to him/herself and gave the repeat command while looking straight ahead. We carried out one such trial per dog and these situations, similarly to the Spontaneous action test , were not practiced during training.

In the above two conditions ( Spontaneous action test and Spontaneous object action test ), although we did not plan to insert a specific delay between the spontaneous action of the dog and the repeat command, the command was issued 5–15 s after the spontaneous action of the dog (this time was needed for the owner to call the dog back from where the action was performed and to give the repeat command).

Spontaneous action control and spontaneous object-action control

In this control condition, after the dog spontaneously performed a well-identifiable action, the owner called it, but did not give the repeat command. Instead, the dog was let free again in the area for 30 s. This was done for both the Spontaneous action test and Spontaneous object-action test , so that every dog participated in two such trials overall.

Different word control and spontaneous action word control

The owner asked the dog to perform a trained action (similarly to the Untrained actions repeat test ). After the dog had performed the action, instead of giving the repeat command, the owner uttered a word of no meaning for the dog (‘Blue’ or ‘Farfalla’), always looking straight ahead.

This control was also carried out in the Spontaneous action test ( Spontaneous action word control ). In this case, after the dog had spontaneously performed an identifiable action, the owner called the dog and said the word of no meaning. This control consisted of two trials for every dog, one with the first action being requested by the owner ( Different word control ) and one in which the owner waited for the dog to perform spontaneously an identifiable action outside of a training/testing context ( Spontaneous action word control ).

Delayed spontaneous tests

The Spontaneous action test and the Spontaneous object-action test were also performed with delays (retention intervals) of 20 s, 1 min, and 1 h between the first identifiable spontaneous action of the dog and the repeat command. The delayed and non-delayed tests were carried out in randomized order.

In the tests with delays of 20 s and 1 min, after the dog had performed a well-identifiable action, the owner called the dog back and walked with it away from the area for the given duration of the delay. When the delay elapsed, the owner walked back to his/her initial position and gave the repeat command. In the tests with 1 h delay, after performing a potentially repeatable action, the dog was placed into its crate by the owner. The dog stayed there for the duration of the delay (all dogs were ‘crate-trained’, i.e., they were accustomed to stay and rest in their crates). This was done to ensure that the only potentially repeatable action performed in that context was the one identified before, so that, being brought in the same context again for the repeat test, the dog would have a possibility to identify and remember it (see also Control for context test). Moreover, taking the dog away from the view of the environment where the action was performed ensured that it could not potentially keep its mind active on the performed action by looking at some environmental stimuli. We carried out one trial per condition/delay (20 s, 1 min, 1 h), per dog.

Control for context

This test was carried out to test if, in the delayed spontaneous tests, the dog recognized and remembered the action previously performed, even if a delay elapsed, by being brought to the same context where it did the action before. We tested a subset of 4 dogs in an identical delayed test as the Spontaneous action test , with an interval of 20 seconds, but giving the repeat command in a different context from the one where they performed the action. After the dog had performed a well-identifiable action in a given area (the living room of the owner), the owner called the dog back and walked with it to a different location (in the garden: N = 3 dogs; in the terrace: N = 1 dog). When 20 seconds from the action previously performed elapsed, the owner gave the repeat command at this different place. We carried out one such trial per dog.

Statistical analysis

Repeat success in the various tests (binary response variable) was analyzed using binomial Generalized Linear Mixed Models (R package ‘lme4’) 46 with dog ID as random term and test condition and/or delay as fixed effects. The effects of explanatory variables were analyzed by likelihood ratio tests (LRT): we provide χ 2 and p-values of likelihood ratio tests of models with and without the explanatory variable.

Informed consent for publication of identifying information/images in an online open-access publication has been acquired 34 .

Ethical statement

All experiments were performed in accordance with relevant guidelines and regulations. The Institutional Committee of Eötvös Loránd University has approved the experiments of this study (N. PE/EA/2021-5/2017).

Data availability

Data will be available upon request.

Mercado, E., Murray, S. O., Uyeyama, R. K., Pack, A. A. & Herman, L. M. Memory for recent actions in the bottlenosed dolphin (Tursiops truncatus): Repetition of arbitrary behaviors using an abstract rule. Anim. Learn. Behav. 26 , 210–218 (1998).

Google Scholar  

Kukzaj, S. A. & Eskelinen, H. C. The “creative dolphin’ revisited: what do dolphins do when asked to vary their behavior. Anim. Behav. Cogn. 1 , 66–77 (2014).

Herman, L. M., Matus, D. S., Herman, E. Y., Ivancic, M. & Pack, A. A. The bottlenose dolphin’s (Tursiops truncatus) understanding of gestures as symbolic representations of its body parts. Learn. Behav. 29 , 250–264 (2001).

Hauser, M. D., Chomsky, N. & Fitch, W. T. The Faculty of Language: What Is It, Who Has It, and How Did It Evolve? Science 298 , 1569–1579 (2002).

ADS   CAS   PubMed   Google Scholar  

Amsterdam, B. K. Mirror self-image reactions before age two. Dev. Psychobiol. 5 , 297–305 (1972).

CAS   PubMed   Google Scholar  

Nielsen, M., Suddendorf, T. & Slaughter, V. Self- recognition beyond the face. Child. Dev. 77 , 176–185 (2006).

PubMed   Google Scholar  

Gallup, G. G. Chimpanzees: self recognition. Science 167 , 86–87 (1970).

ADS   Google Scholar  

Foote, A. L. & Crystal, J. D. Metacognition in the rat. Curr. Biol. 17 , 551–555 (2007).

CAS   PubMed   PubMed Central   Google Scholar  

Herman, L. M. Body and self in dolphins. Conscious. Cogn. 21 , 526–545 (2012).

Plotnik, J. M., de Waal, F. B. M. & Reiss, D. Self-recognition in an Asian elephant. Proc. Natl. Acad. Sci. USA 103 , 17053–17057 (2006).

Reiss, D. & Marino, L. Mirror self-recognition in the bottlenose dolphin: A case of cognitive convergence. Proc. Natl. Acad. Sci. U.S.A. 98 , 5937–5942 (2001).

ADS   CAS   PubMed   PubMed Central   Google Scholar  

Prior, H., Schwarz, A. & Güntürkün, O. Mirror-induced behavior in the magpie (Pica pica): Evidence of self-recognition. PloS. Biol. 6 , 1642–1650 (2008).

CAS   Google Scholar  

Kohda, M. et al . If a fish can pass the mark test, what are the implications for consciousness and self-awareness testing in animals?. PLoS biology , 17 (2019).

Heyes, C. M. Reflections on self-recognition in primates. Anim. Behav. 47 , 909–919 (1994).

Rochat, P. Five levels of self-awareness as they unfold early in life. Conscious. Cogn. 12 , 717–731 (2003).

Platek, S. M., Thomoson, J. W., Gordon, G. Jr. & Gallup, Jr. Cross-modal self-recognition: the role of visual, auditory, and olfactory primes. Conscious. Cogn . 13 , 197–210 (2004).

Gallup, G. G. The nature of visual self-recognition revisited. Trends. Cogn. Sci. 18 , 57–58 (2014).

Howell, J. T., Toukhsati, S., Conduit, R. & Bennett, P. Do dogs use a mirror to find hidden food? J. Vet. Behav. 8 , 425–430 (2013).

Lenkei, R., Faragó, T., Kovács, D., Zsilák, B. & Pongrácz, P. That dog won’t fit: body size awareness in dogs. Anim. Cogn. 23 , 337–350 (2020).

Horowitz, A. Smelling themselves: Dogs investigate their own odours longer when modified in an “olfactory mirror” test. Behav. Proc. 143 , 17–24 (2017).

Miklósi, Á. & Topál, J. What does it take to become ‘best friends’? Evolutionary changes in canine social competence. Trends. Cogn. Sci. 17 , 287–294 (2014).

Fugazza, C., Pogány, Á. & Miklósi, Á. Recall of others’ actions after incidental encoding reveals episodic-like memory in dogs. Curr. Biol. 26 , 3209–3213 (2016).

Tulving, E. How many memory systems are there. Am. Psychol. 40 , 385–398 (1985).

Suddendorf, T. & Corballis, M. C. Behavioural evidence for mental time travel in nonhuman animals. Behav. Brain. Res. 215 , 292–298 (2010).

Cheng, S., Werning, M. & Suddendrof, T. Dissociating memory traces and scenario construction in mental time travel. Neurosci. Biobehav. Rev. 60 , 82–89 (2016).

Tulving, E. Episodic and semantic memory. In E. Tulving and W. Donaldson (Eds.), Organization of memory (pp. 381–403). San Diego, CA: Academic Press. (1972).

Panoz-Brown, D. et al . Replay of episodic memories in the Rat. Curr. Biol. 28 , 1–7 (2018).

Zentall, T. R., Singer, R. A. & Stagner, J. P. Episodic-like memory: pigeons can report location pecked when unexpectedly asked. Behav. Proc. 79 , 93–98 (2008).

Clayton, N. S. & Dickinson, A. Episodic-like memory during cache-recovery by scrub jays. Nature. 395 , 272–274 (1998).

Kaminski, J., Fisher, J. & Call, J. Prospective object search in dogs: mixed evidence for knowledge of What and Where. Anim. Cogn. 11 , 367–371 (2008).

Martin-Ordas, G., Bernsten, D. & Call, J. Memory for Distant Past Events in Chimpanzees and Orangutans. Curr. Biol. 23 , 1438–1441 (2013).

Kart-Teke, E., De Souza Silva, M. A., Huston, J. P. & Dere, E. Wistar rats show episodic-like memory for unique experiences. Neurobiol. Learn. Mem. 85 , 173–182 (2006).

Zentall, T. R. Animals represent the past and the future. Evol. Psychol. 11 , 573–590 (2013).

Crystal, J. D. Comparative cogntion: Action imitation using Episodic memory. Curr. Biol. 26 , 1226–1228 (2016).

Zohu, W., Hohmann, A. G. & Crystal, J. D. Rats answer an unexpected question after incidental encoding. Curr. Biol. 22 , 1149–1153 (2012).

Shea, N. & Heyes, C. Metamemory as evidence of animal consciousness: The type that does the trick. Biol. Philos. 25 , 95–110 (2010).

Tulving, E. & Schacter, D. L. Priming and Human Memory Systems. Science 247 , 4940 (1990).

Fugazza, C. & Miklósi, Á. Deferred imitation and declarative memory in domestic dogs. Anim. Cogn. 17 , 237–247 (2014).

Talamini, L. M. & Gorree, E. Aging memories: differential decay of episodic memory components. Learn. Mem. 19 , 239–246 (2012).

Hayes, S. M., Nadel, L. & Ryan, L. The Effect of Scene Context on Episodic Object Recognition: Parahippocampal Cortex Mediates Memory Encoding and Retrieval Success. Hippocampus 17 , 873–889 (2007).

PubMed   PubMed Central   Google Scholar  

Hollingworth, A. Object-Position Binding, A. in Visual Memory for Natural Scenes and Object Arrays. J. Exp. Psychol. Hum. Percept. Perform. 33 , 31–47 (2007).

Conway, M. A. Memory and the self. J. Mem. Lang. 53 , 594–628 (2005).

Suddendorf, T. & Corballis, M. C. The evolution of foresight: What is mental time travel, and is it unique to humans? Behav. Brain. Sci. 30 , 299–313 (2007).

Miklósi, Á. Dog behaviour, evolution and cognition. Second edition, Oxford university press (2015).

Topal, J., Byrne, R., Miklosi, A. & Csanyi, V. Reproducing human actions and action sequences: “Do as I Do!” in a dog. Anim Cogn 9 , 355–367 (2006).

Bates, D., Maechler, M., Bolker, B. & Walker, S. Fitting Linear Mixed-Effects Models Using lme4. J. Stat. Softw. 67 , 1–48 (2015).

Download references

Acknowledgements

This study was supported by the National Brain Research Program (2017-1.2.1-NKP-2017-00002). A.M. received funding from MTA-ELTE Comparative Ethology Research Group (MTA01 031), and from the ELTE Institutional Excellence Program supported by the National Research, Development and Innovation Office (NKFIH-1157-8/2019-DT). A.P. was supported by the János Bolyai Research Scholarship of the Hungarian Academy of Sciences and by the ÚNKP-19-4 New National Excellence Program of the Ministry of Innovation and Technology, Hungary. We are extremely grateful to Jena Olio, Marco Ojeda, Jorge Isaac Galván, Laura Carmignani, Paola Roveda and Angela Riccardi for participating in this study with their dogs, to Marco Ojeda for directing and editing the video abstract and to Sean O’Hara for registering the voice over of the video abstract.

Author information

Authors and affiliations.

Department of Ethology, Eötvös Loránd University, Budapest, Hungary

Claudia Fugazza, Péter Pongrácz, Ákos Pogány, Rita Lenkei & Ádám Miklósi

MTA-ELTE Comparative Ethology Research Group, Budapest, Hungary

• Ádám Miklósi

You can also search for this author in PubMed   Google Scholar

Contributions

C.F. and P.P. conceived the study, C.F. devised the experiments and all authors contributed in designing the experimental protocol, C.F. collected data, A.P. analysed the results, C.F. drafted the manuscript and C.F., P.P., A.P., R.L. and A.M. revised it.

Corresponding author

Correspondence to Claudia Fugazza .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary material., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Fugazza, C., Pongrácz, P., Pogány, Á. et al. Mental representation and episodic-like memory of own actions in dogs. Sci Rep 10 , 10449 (2020). https://doi.org/10.1038/s41598-020-67302-0

Download citation

Received : 10 July 2019

Accepted : 05 June 2020

Published : 26 June 2020

DOI : https://doi.org/10.1038/s41598-020-67302-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

Body size awareness matters when dogs decide whether to detour an obstacle or opt for a shortcut.

• Péter Pongrácz
• Petra Dobos
• Rita Lenkei

Scientific Reports (2023)

Investigating responses to object-labels in the domestic dog (Canis familiaris)

• Hanna Kőszegi
• Claudia Fugazza
• Attila Andics

Second verse, same as the first: learning generalizable relational concepts through functional repetition

• Eduardo Mercado
• Allison Scagel

Animal Cognition (2023)

• Sara Torres Ortiz
• Simeon Q. Smeele
• Auguste M. P. von Bayern

Scientific Reports (2022)

Word learning dogs (Canis familiaris) provide an animal model for studying exceptional performance

Scientific Reports (2021)

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

• Reference Manager
• Simple TEXT file

People also looked at

Conceptual analysis article, dogs supporting human health and well-being: a biopsychosocial approach.

• 1 Department of Psychiatry, Center for Human Animal Interaction, School of Medicine, Virginia Commonwealth University, Richmond, VA, United States
• 2 Human-Animal Bond in Colorado, School of Social Work, Colorado State University, Fort Collins, CO, United States
• 3 Department of Education, California State Polytechnic University, Pomona, CA, United States
• 4 Division of Social Sciences and Natural Sciences, Seaver College, Pepperdine University, Malibu, CA, United States

Humans have long realized that dogs can be helpful, in a number of ways, to achieving important goals. This is evident from our earliest interactions involving the shared goal of avoiding predators and acquiring food, to our more recent inclusion of dogs in a variety of contexts including therapeutic and educational settings. This paper utilizes a longstanding theoretical framework- the biopsychosocial model- to contextualize the existing research on a broad spectrum of settings and populations in which dogs have been included as an adjunct or complementary therapy to improve some aspect of human health and well-being. A wide variety of evidence is considered within key topical areas including cognition, learning disorders, neurotypical and neurodiverse populations, mental and physical health, and disabilities. A dynamic version of the biopsychosocial model is used to organize and discuss the findings, to consider how possible mechanisms of action may impact overall human health and well-being, and to frame and guide future research questions and investigations.

Introduction – A Historical Perspective on Dog-Human Relationships

The modern relationship between humans and dogs is undoubtedly unique. With a shared evolutionary history spanning tens of thousands of years ( 1 ), dogs have filled a unique niche in our lives as man's best friend. Through the processes of domestication and natural selection, dogs have become adept at socializing with humans. For example, research suggests dogs are sensitive to our emotional states ( 2 ) as well as our social gestures ( 3 ), and they also can communicate with us using complex cues such as gaze alternation ( 4 ). In addition, dogs can form complex attachment relationships with humans that mirror that of infant-caregiver relationships ( 5 ).

In today's society, dog companionship is widely prevalent worldwide. In the United States, 63 million households have a pet dog, a majority of which consider their dog a member of their family ( 6 ). In addition to living in our homes, dogs have also become increasingly widespread in applications to assist individuals with disabilities as assistance dogs. During and following World War I, formal training of dogs as assistance animals began particularly for individuals with visual impairments in Germany and the United States ( 7 ). Following World War II, formal training for other roles, such as mobility and hearing assistance, started to increase in prevalence. Over the decades, the roles of assistance dogs have expanded to assist numerous disabilities and conditions including medical conditions such as epilepsy and diabetes and mental health disorders such as posttraumatic stress disorder (PTSD). At the same time, society has also seen increasing applications of dogs incorporated into working roles including detection, hunting, herding, and protection ( 8 , 9 ).

In addition to these working roles, dogs have also been instrumental in supporting humans in other therapeutic ways. In the early 1960s, animal-assisted interventions (AAI) began to evolve with the pioneering work of Boris Levinson, Elizabeth O'Leary Corson, and Samuel Corson. Levinson, a child psychologist practicing since the 1950s, noticed a child who was nonverbal and withdrawn during therapy began interacting with his dog, Jingles, in an unplanned interaction. This experience caused Levinson to begin his pioneering work in creating the foundations for AAI as an adjunct to treatment ( 10 ). In the 1970s, Samuel Corson and Elizabeth O'Leary Corson were some of the first researchers to empirically study canine-assisted interventions. Like Levinson, they inadvertently discovered that some of their patients with psychiatric disorders were interested in the dogs and that their patients with psychiatric disorders communicated more easily with each other and the staff when in the company of the dogs ( 11 , 12 ). Over the following decades, therapy dogs have been increasingly found to provide support for individuals with diverse needs in a wide array of settings ( 13 ).

Theoretical Framework for Dog Interaction Benefits

For over 40 years, the biopsychosocial model ( 14 ) has been widely used to conceptualize how biological, psychological, and social influences combine to determine human health and well-being. Biological influences refer to physiological changes such as blood pressure, cortisol, and heart rate, among others; psychological influences include personality, mood, and emotions, among others; and social influences refer to cultural, socio-economic, social relationships with others, family dynamics, and related matters. Figure 1 presents a graphical illustration of the relationship among these three influences in determining overall health and well-being. Although the model has dominated research and theory in health psychology for decades, more recently, it was re-envisioned as a more dynamic system ( 15 ) that construes human health as the result of the reciprocal influences of biological, psychological and social factors that unfold over personal and historical time. For example, if a person breaks his/her arm, there will be a biological impact in that immune and muscle systems respond and compensate. Social, or interpersonal, changes may occur when support or assistance is offered by others. Psychological changes will occur as a result of adjusting to and coping with the injury. Thus, the injury represents a dynamic influence initiated at one point in time and extending forward in time with diminishing impact as healing occurs.

Figure 1 . A biopsychosocial perspective of how biological, psychological, and social influences may impact one another (solid lined arrows) and influence human health and well-being (represented here by the large thick circular shape).

This dynamic biopsychosocial approach to understanding health and well-being is appealing to the field of human-animal interaction (HAI) because of the dynamic nature of the relationship between humans and animals. For example, a person may acquire many dogs over his/her lifetime, perhaps from childhood to old age, and each of those dogs may sequentially develop from puppyhood to old age in that time. Behaviorally, the way the human and the dog interact is likely to be different across the lifespans of both species. From a biopsychosocial model perspective, the dynamic nature of the human-canine relationship may differentially interact with each of the three influencers (biological, psychological, and social) of human health and well-being over the trajectories of both beings. Notably, these influencers are not fixed, but rather have an interactional effect with each other over time.

While a person's biological, psychological, and social health may affect the relationship between that person and dogs with whom interactions occur, the focus of this manuscript is on the reverse: how owning or interacting with a dog may impact each of the psychological, biological, and social influencers of human health. We will also present relevant research and discuss potential mechanisms by which dogs may, or may not, contribute to human health and well-being according to the biopsychosocial model. Finally, we will emphasize how the biopsychosocial theory can be easily utilized to provide firmer theoretical foundations for future HAI research and applications to therapeutic practice and daily life.

Psychological Influences

Much research has been conducted on the impact of dog ownership and dog interactions on human psychological health and functioning. Frequent interactions with a dog, either through ownership or through long-term interventions, have been associated with positive psychological outcomes across the lifespan [for a systematic review of this evidence see ( 16 )]. One psychological aspect of interest to many HAI researchers is depression, especially among older adults. However, the relationship of pet dog ownership and depression over the lifespan continues to have inconsistent and inconclusive findings ( 16 ). Nevertheless, there are examples in the literature highlighting the beneficial role of dog ownership in reducing depression. As is frequently the case in HAI, the evidence from intervention studies is stronger than that of pet ownership studies ( 16 ), with the preponderance of this evidence linking animal-assisted interventions to a decrease in depression, as measured by self-report indices. Among the mechanisms for this reduction in depression are biological and social influences. For example, one such study found that an attachment relationship with a pet dog may serve as a coping resource for older women by buffering the relationship between loneliness (also measured by self-report indices) and depression, such that the presence of the pet dog appears to ameliorate the potential for loneliness to exacerbate depression ( 17 ). A causal relationship between dog ownership and mental health is difficult to determine. Not only may owning a pet dog increase stress, but those who are already suffering from loneliness or depression may be more inclined to have a pet dog than those who do not.

Another psychological outcome related to dog interaction that receives considerable research attention is anxiety. Studies have found that short-term, unstructured interactions with a therapy dog can significantly reduce self-reported anxiety and distress levels [e.g., ( 18 )]. For example, children with their pet dog or a therapy dog present during a stressful task exhibit lower perceived stress and more positive affect compared to when alone ( 19 ), when a parent was present ( 20 ), or when a stuffed dog was present ( 21 ). In addition to psychological mechanisms, there are social and biological mechanisms at play as well. In these short-term stressful contexts, a dog may serve as both a comforting, nonjudgmental presence as well as a positive tactile and sensory distraction. Dog interaction might also reduce anxiety and distress by influencing emotion regulation while coping with a stressor ( 22 ). During animal-assisted therapy, having a dog present during psychotherapy such as cognitive behavioral therapy can aid in decreasing self-reported anxious arousal and distress for patients who have experienced trauma, making the therapeutic treatment process more effective ( 23 ).

In addition to the negative aspects of psychological functioning, HAI research has also aimed to quantify the effects of dog interaction and ownership on positive psychological experiences such as happiness and well-being. Some studies have found that dog ownership is associated with higher life satisfaction and greater well-being ( 24 ), while other studies show that this is the case only when the dog provided social support ( 25 ) or satisfied the owner's needs ( 26 ). However, other large-scale surveys have found no significant differences in self-reported happiness between dog owners, cat owners, and non-pet owners ( 27 ), contributing to mixed findings. Recent discussions argue that too much focus has been placed on the relationship between mental health and the simple variable of dog ownership, when the specific activities that owners engage in with their dogs (e.g., walking, tactile interaction, and shared activities,) may be more important in explaining positive well-being ( 28 ). Further, many other factors may be driving these inconsistent findings in depression, anxiety, and well-being, including the owner's personality ( 24 ), gender and marital status ( 29 ), and attachment to the dog ( 30 ).

Dogs may also provide a source of motivation; for example, people with dogs are more likely to comply with the rigors of their daily life ( 31 ). The relationship with a pet dog may provide motivation to do things that may be less desirable. For example, for older adults who own pets, it is not uncommon for them to be more involved in daily life activities because of the need to take care of their animals ( 32 ). Likewise, children also complete less desired activities due to their relationship with the dog [for a discussion of this topic see ( 33 )].

An accumulation of research also suggests that dog interaction may have specific psychological benefits for individuals with physical disabilities and chronic conditions. Cohabitating with a specially trained assistance dog, including guide, hearing, and service dogs, can be associated with increased psychological and emotional functioning among individuals with disabilities ( 34 ). For individuals with mental disorders such as posttraumatic stress disorder (PTSD), recent research has also found that having a psychiatric service dog is associated with fewer PTSD symptoms, less depression and anxiety, and better quality of life [For a review see ( 35 )]. These benefits appear to be due to a combination of the service dog's specific trained tasks and aspects inherent to cohabitating with a pet dog, including having a source of love, nonjudgmental social support, and companionship ( 36 ).

Similar research has also highlighted the value of dogs for children with disorders of executive functioning and self-regulation, especially autism spectrum disorder (ASD) and attention-deficit/hyperactivity disorder (ADHD). For some children with ASD, dogs may provide a calming and positive presence ( 37 ) and may both reduce anxiety ( 38 ) and improve problematic behaviors ( 39 ). Parents report that both pet dogs and service dogs can provide certain benefits for children with ASD, including benefits to children's moods, sleep, and behavior ( 40 , 41 ). Therapy dogs have also been found to be impactful in supporting children with ADHD in their emotional regulation ( 42 ) and aspects of character development ( 43 ). Nevertheless, the outcome of dog interactions may not be positive for all individuals with ASD and ADHD; despite evidence of psychological benefits of dog interaction for some children, others may be fearful or become over-stimulated by dogs ( 44 ).

In addition to impacts on psychological health, dog interaction can also impact psychological functioning, cognition, and learning. Among children, emerging research suggests short-term interactions with a therapy dog may lead to improvements in specific aspects of learning and cognition. A recent systematic review of research on therapy dog reading programs indicated that reading to a dog has a number of beneficial effects including improved reading performance ( 45 ). Studies suggest that interacting with a therapy dog may also improve speed and accuracy on cognitive (e.g., memory, categorization, adherence to instructions) and motor skills tasks among preschool-aged children compared to interacting with a stuffed dog or human ( 46 ). Similarly, a recent study showed that 10–14-year-old children had greater frontal lobe activity in the presence of a real dog as compared to a robotic dog, indicating a higher level of neuropsychological attention ( 47 ).

Among young adults, similar effects on cognition and learning have been found. Numerous colleges and universities now offer interactions with therapy dogs, typically during high stress times (such as before exams). In this sense, a biological mechanism through which dog interaction may positively impact cognition and learning is via stress reduction and improvement in positive affect. Even such short and infrequent interactions with therapy dogs may decrease perceived stress and increase perceived happiness in college students [e.g., ( 48 , 49 )]. Further, some institutions have permanent resident therapy dogs and/or long-term intervention programs; one such program showed that students who interacted with therapy dogs for 8 weeks reported significantly less homesickness and greater satisfaction with life than wait-listed controls ( 50 ). These effects may translate to additional effects on students' academic success, learning, and cognition. For instance, a recent randomized controlled trial ( 51 ) paired a standard academic stress management program with therapy dog interaction; the pairing produced significantly higher levels of self-reported enjoyment, usefulness, self-regulation, and behavior change than the stress management program or dog interaction alone. However, when therapy dog interaction is closely paired with more specific learning experiences, beneficial effects on stress remain, but benefits to academic performance may not manifest. For example, a recent study showed that interacting with a therapy dog resulted in significant improvements in students' perceived stress and mood, but not in actual exam scores ( 52 ). Similarly, interacting with a therapy dog during the learning and recall phase of a memory test did not improve memory compared to a control group ( 53 ). Taken together, dog interaction may improve stress and affect among college-aged adults as well as dimensions important for academic success and learning, but these results may or may not translate to cognitive performance benefits.

Biological Influences

The psychological and biological effects of HAI are often closely interwoven, as seen in the Psychological Influences section above and as demonstrated by the frequency with which psychological effects are evaluated using biological assessments of stress, anxiety, and arousal ( 54 ). For example, a plethora of studies have examined how short-term interactions with dogs can influence stress by measuring physiological biomarkers. Studies have found that dog interaction can influence parameters such as blood pressure, heart rate, and electrodermal activity ( 55 ) as well as neurochemical indicators of affiliative behavior [e.g., beta-endorphins, prolactin, and dopamine; ( 56 )].

However, one of the most popular physiological measures in HAI research is the stress hormone cortisol ( 57 ). Studies have found that short-term interactions with a dog can decrease both subjective stress and circulating cortisol concentrations [e.g., ( 58 )]. Cohabitating with a dog has also been found to impact circulating cortisol after waking among children with ASD ( 39 ) and military veterans with PTSD ( 59 ). Experimental studies have also examined how having a dog present may modulate the stress response and cortisol secretion among individuals undergoing a stressful situation. Among adults, studies have found that having a dog present during a socially stressful paradigm can attenuate cortisol compared to when alone or with a human friend ( 60 ). A recent randomized controlled trial similarly found that interacting with a therapy dog, for 20 min, two times per week, over a 4-week period resulted in reduced cortisol (basal and diurnal measurement) among typically developing and special needs school children compared to the same duration and length of delivery for a yoga relaxation or a classroom as usual control group ( 61 ). However, it is of note that many methodologically rigorous studies have not found significant effects of interacting with a dog on physiological parameters, including salivary cortisol ( 21 , 62 , 63 ). A recent review of salivary bioscience research in human-animal interaction concluded that significant variation exists with regards to sampling paradigms, storage and assaying methods, and analytic strategies, contributing to variation in findings across the field ( 57 ).

As research quantifying the physiological outcomes from dog interaction continues to increase, so does research attempting to understand the underlying mechanisms of action leading to stress reduction. One theoretical rationale for dogs' stress-reducing benefits consists of the dog's ability to provide non-judgmental social support ( 60 ), improve positive affect ( 64 ), and provide a calming presence ( 22 ). Dogs may also contribute to a feeling of perceived safety and provide a tactile and grounding comfort ( 65 ). For these reasons, dogs are often incorporated into treatment and recovery for individuals who have experienced a traumatic event ( 66 ). Another mechanism contributing to these stress reducing benefits may be tactile stimulation and distraction derived from petting or stroking a dog. For example, Beetz et al. ( 67 ) found that the more time a child spent stroking the dog before a stressful task, the larger the magnitude of cortisol decrease. In fact, calming tactile interactions such as stroking, touching, and petting may be a key mechanism explaining animal-specific benefits to stress physiology, as touch is more socially appropriate in interactions with animals than as with other people ( 22 ). While there are many hypothesized mechanisms underlying positive psychophysiological change following human-dog interaction, more research is needed to determine how individual differences in humans, animals, and the human-animal relationship affects outcomes ( 21 , 57 , 62 , 63 ).

Another mechanism in which positive dog interaction may result in psychophysiological benefits is via the secretion of oxytocin. Oxytocin not only buffers the stress response and cortisol secretion ( 68 ) but is also involved emotion, trust, and bonding ( 69 ). The oxytocin system has been hypothesized to be a primary mechanistic pathway involved in human-dog interactions ( 70 ). Positive dog-owner interactions including stroking, petting, and talking have been shown to result in increased oxytocin levels in both dog owners and dogs, which has been related to the strength of the owner-dog relationship ( 71 ) and dog-human affiliative behaviors ( 72 , 73 ). Some studies have also found differential effects in oxytocin reactivity after dog interaction between human males and females ( 74 ), giving context to potential gender and/or hormonal differences in dog-human interactions. However, even though the oxytocin system exhibits potential as a pathway by which dogs provide psychophysiological benefits, it should be noted that mixed findings and methodological and measurement differences limit strong conclusions ( 75 ).

In regards to pet dog ownership, many studies have also sought to understand the biological effects of long-term interactions with a pet dog. Some research suggests that sharing animal-associated microbes with a pet dog can have long-term impacts on human health ( 76 ) while others have found that cohabitating with a pet dog can be beneficial for child allergies ( 77 ) and immune system development ( 78 ). However, most research on the long-term health impacts of pet dog ownership has focused on cardiovascular functioning. Epidemiological research suggests that dog ownership is linked to greater physical activity levels (presumably linked to dog-walking), and reduced risk for cardiovascular disease, stroke, and all-cause mortality [for a summary see ( 79 )]. A recent meta-analysis of ten studies amassing data from over three million participants found that pet dog ownership was associated with a 31% risk reduction for mortality due to cardiovascular disease ( 80 ). However, dog ownership research of this nature will always suffer from an important chicken and egg type question: do dogs make us healthier, or do healthy people opt to own dogs?

Social Influences

A final way in which dog companionship and interaction may contribute to human health and well-being is through the social realm. Dogs may impact social functioning by providing direct social support ( 81 ) and a source of an attachment bond ( 82 ) which in turn may contribute to better social and mental health by providing companionship. Acquiring a pet dog has been reported to reduce both short-term and long-term self-reported loneliness ( 83 ). Particularly for those who live alone, dog ownership may serve as a protective factor against loneliness in times of social isolation, such as during the COVID-19 pandemic ( 84 ). Among older adults living in long-term care facilities or who live alone, dog visitation may also decrease loneliness by providing a source of meaningful companionship and social connectedness ( 85 , 86 ). However, the literature on pet dogs and loneliness is also characterized by mixed findings, raising the possibility that dog ownership may be a response to loneliness rather than protection from loneliness. Further, there remains a lack of high quality research in this area which limits any causal conclusions ( 87 ).

Another way in which the social support from a pet dog may benefit social functioning is by facilitating social interactions with others. For example, observational studies have found that being accompanied by a dog in public increases the frequency of received social interactions ( 88 ) and social acknowledgments [e.g., friendly glances, smiles; ( 89 )]. For those who engage in dog walking, social interactions are perceived as a rewarding side effect ( 90 ). Dogs can also provide a source of social capital, defined as the glue that holds society together ( 91 ). The research of Wood and colleagues ( 92 ) suggests that dogs can function as facilitators for social contact and interaction, with pet owners reporting higher perceptions of suburb friendliness and more social interactions with neighbors compared to non-pet owners.

For children and adolescents, pet dog ownership may contribute to healthy social development. Positive child–pet dog interactions have been shown to have benefits to children's social competence, interactions, and play behavior [for a review see ( 93 )]. Not only can children form attachment relationships with dogs ( 94 ), but pet dogs may promote feelings of safety and security ( 95 ) that can facilitate childhood social development. Pet ownership may also help children develop skills to form and maintain social relationships with their peers ( 96 ). For example, cross-sectional studies found that children with a pet dog in the home have fewer peer problems and have more prosocial behavior with children without a dog [e.g., ( 97 , 98 )].

Among children with developmental disorders, dog interaction has also been similarly shown to impact social functioning. For children with ADHD, two randomized controlled trials have found that 12 weeks of visits with a therapy dog, incorporated into curricula designed to improve skills and reduce behavioral problems, can result in improved social skills, prosocial behaviors, and perceptions of social competence ( 42 , 43 ). One potential explanation for these benefits is that children may interpret the dogs' nonverbal communication as less threatening and easier to interpret than human interaction ( 99 , 100 ). A recent eye-tracking study found that children with ASD exhibit a bias in social attention to animal faces, including dogs, compared to human faces ( 101 ). The presence of a dog in clinical applications may also promote more social engagement with a therapist while reducing negative behaviors ( 102 , 103 ). Further, there is some evidence that having a pet dog in the home can have a positive impact on social interactions of children with ASD, especially among verbal children, while teaching children responsibility and empathetic behavior ( 104 , 105 ).

Potential Mechanisms of Action

We have discussed how, in the psychological realm, interacting with a dog can positively relate to depression, anxiety, and well-being as well as psychological functioning in the areas of cognition, learning, and attention. It is interesting to note that most psychological constructs are measured using self-report indices, such as the Beck Depression Inventory ( 106 ) or the UCLA Loneliness Scale ( 107 ), while a smaller group of constructs are measured using speed and accuracy to detect targets (attention) or to remember information (learning and memory). In the biological realm, we discussed how interacting with dogs can influence stress-related physiological parameters and long-term biological and cardiovascular health. Biological measures are often recorded in real-time, such as heart rate or blood pressure, or are collected at critical time points during the study (e.g., saliva, urine, or blood samples for such measures as cortisol or oxytocin). Finally, we discussed the social realm, in which interacting with a dog can provide social support, facilitate social interactions, and improve social development and social skills. Measures used to assess variables in the social realm include self-report indices (e.g., demographics such as marital status, numbers of family members and friends), real time observations of social interactions (e.g., video analyses of interactions using ethograms), and parent/teacher reports of social functioning [e.g., Social Skills Rating System; ( 108 )]. To better understand and organize these various findings, we now consider potential mechanisms of action in the context of the biopsychosocial model, and as part of this discussion we will consider the potential for different types of measurement to have their own influence.

The mechanisms that underly positive human-dog interactions are likely to be interrelated and broadly, yet differentially, impactful across the three influencers of health (biological, psychological and social). According to the biopsychosocial model, impacts on one of the influencers of health is likely to impact the others ( 14 ). Further, an underlying mechanism of change may have a larger immediate impact on one realm than on the other two ( 15 ). Although this applies to the many influences we have discussed above, we will describe a reduction in stress as a more detailed example of how the biopsychosocial model can be considered. Stress is likely to have an immediate and measurable impact on the biological system through endocrinological (e.g., changes in cortisol) and psychophysiological (e.g., changes in blood pressure) processes. This same reduction in stress is likely to impact the psychological system through changes in mood or affect, concentration, and motivation, but that impact may not be immediately measurable or may be smaller in magnitude. This conjectured delay or reduction in effect size stems at least in part, from the way these changes are typically measured and the time course for potential effects to become measurable. For example, some biological changes indicative of increased stress (e.g., heart rate) can be measured in direct correspondence with the experimental manipulations (e.g., interacting with the dog vs. experiencing a control condition), and provide real time biological indications of changes in stress levels. Psychological indications of stress may be measured by a self-report survey instrument assessing state or trait anxiety. This type of measure cannot be completed in real time during the various experimental conditions (e.g., interacting with the dog vs. experiencing a control condition), but must be completed at some point following the experimental manipulation. It is possible that psychological measures are not as immediately sensitive to changes in the constructs they measure because of the required delay between manipulation and measurement. Such a delay may underestimate the real time effect as it may fade over time. Finally, reductions in stress have the potential to impact social systems by increasing social approaches and acceptance of approaches by others, but that impact may be of a small size or require even more time to be measurable. For example, exposure to stress may have immediate physiological effects, but it could take more time (prolonged exposure to stress) for those effects to impact some measures of social influence such as number of friends.

In Figure 2 , the mechanism of stress reduction is used as one example for the purposes of this discussion to exemplify how human-dog interactions may influence human health and well-being, as explained by the biopsychosocial model. Stress reduction may have a more immediate or larger impact on the biological realm as demonstrated by the larger arrow, while having a smaller (or perhaps delayed) impact on the psychological realm and an even smaller (or potentially more delayed) impact on the social realm.

Figure 2 . An example of the potential for differential impact (represented by the different arrow thickness) of one mechanism of action (stress reduction) on the three realms of influence of overall health and well-being (depicted by the larger encompassing circle).

Based on the research described earlier, we have seen that interacting with a dog can have stress reducing impacts in the biological realm such as decreased cortisol, heart rate, and blood pressure, and increases in oxytocin. In the psychological realm, stress reduction can be a driver of immediate improvements in self-report measures of stress, mood, and anxiety and more delayed improvements in overall mental health and quality of life. The social realm is also likely to be directly and indirectly impacted by this stress reduction from both immediate and delayed psychophysiological changes as well as more long-term improvements in social support, social networks, social development, and overall social health. Therefore, it is important to consider the dynamic nature of these three realms in that there may be a strong immediate effect of dog interaction on one realm, but a lesser, delayed impact in the other two realms. Similar to our more detailed example of stress above, other influences we have discussed (e.g., social support, positive affect, etc.) are likewise mechanisms that operate in a similar reciprocal biopsychosocial framework. Further, although it likely that the three influences are interrelated, it is not known from the current evidence the degree to which they may be interrelated and thus have shared and overlapping effects on one another and on overall health and well-being. Therefore, a consideration of mechanisms that influence human-dog interactions from a dynamic and flexible biopsychosocial perspective, instead of from a single realm, is an important addition to the study of human-animal interaction.

Conclusion and Future Directions

In conclusion, the biopsychosocial model is a promising theoretical model to be applied to human-animal interaction research for several reasons. First, the field of HAI has been plagued by mixed findings in which some research suggests that dogs have beneficial effects on human health and well-being and others suggest no effect or even a negative effect [for a discussion see ( 109 )]. This variability in HAI research outcomes caused by differing methodologies, measurement, populations, and interventions is described in detail by Rodriguez et al. ( 110 ). However, we also argue that some of the variability seen in HAI research may be explained by the potential for differential immediate and delayed impacts within each of the three biopsychosocial model realms. For example, if dog interaction shows immediate reduction in physiological measures of stress, how long does that reduction last, and do we see corresponding immediate and/or delayed responses in the psychological and social realms? Therefore, more information about differential impacts of dog interactions on each of the three influencers at various points in time is needed. In addition, it may be necessary to apply a variety of measures (at least one measure per influencer realm) over time to fully disentangle the existing mixed results in the field of HAI.

Secondly, due to the flexibility that this dynamic biopsychosocial model offers in explaining HAI research outcomes, we propose this model as an effective avenue to promote future theoretically grounded research in our field. Saleh ( 111 ) stresses that practice, research, and theory are the corner stones of any field, HAI is not exempt from this consideration. The field of HAI will benefit from applying an accepted model, like the biopsychosocial model, because it provides a useful framework for understanding and predicting how interactions between humans and animals impacts human health and well-being. As Saleh ( 111 ) explains, “it is the result of the relationship between the process of inquiry (research) and the product of knowledge (theory)” that our understanding of a process may become clearer. Therefore, current research should continue to modify and impact a present theory, which should act as a guide for researchers to constantly generate and test the basis of a theory ( 111 ). The findings from such theory-driven research could then help practitioners, as well as health care policy makers, in how to effectively incorporate dogs in therapeutic settings and in homes.

Lastly, the reciprocal relationship of the psychological, biological, and social domains can be used to elucidate the mechanisms that both impact and are impacted by interactions between humans and animals. Theory-driven science (for which we have proposed the biopsychosocial model as a useful framework) should be used to influence and inform research, practice, and policy. Thus, researchers and practitioners applying the biopsychosocial model will be instrumental not only in guiding future research in the field, but also in clarifying existing research as well people's perceptions of benefits derived from canine-human interactions.

Author Contributions

NG provided the initial organization and theoretical framework. All authors wrote and edited the document in shared collaboration and discussed and conceived the idea for the paper.

As part of the conferment of Fellowship status to all authors, the Wallis Annenberg Petspace provided the funding for publication fees of this document.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

The authors wish to express their appreciation to the Wallis Annenberg Petspace for supporting this theoretical framework and exploration of the Human-Canine bond.

1. Thalmann O, Shapiro B, Cui P, Schuenemann VJ, Sawyer SK, Greenfield D, et al. Complete mitochondrial genomes of ancient canids suggest a European origin of domestic dogs. Science. (2013) 342:871–4. doi: 10.1126/science.1243650

PubMed Abstract | CrossRef Full Text | Google Scholar

2. Albuquerque N, Guo K, Wilkinson A, Savalli C, Otta E, Mills D. Dogs recognize dog and human emotions. Biol Lett. (2016) 12:20150883. doi: 10.1098/rsbl.2015.0883

3. Hare B, Brown M, Williamson C, Tomasello M. The domestication of social cognition in dogs. Science. (2002) 298:1634–6. doi: 10.1126/science.1072702

4. Miklósi Á, Kubinyi E, Topál J, Gácsi M, Virányi Z, Csányi V. A simple reason for a big difference: wolves do not look back at humans, but dogs do. Curr Biol. (2003) 13:763–6. doi: 10.1016/S0960-9822(03)00263-X

5. Payne E, Bennett PC, Mcgreevy PD. Current perspectives on attachment and bonding in the dog–human dyad. Psychol Res Behav Manag. (2015) 8:71. doi: 10.2147/PRBM.S74972

6. American Pet Products Association. 2019–2020 APPA National Pet Owners Survey . Stamford, CT: American Pet Products Association (2019).

Google Scholar

7. Fishman GA. When your eyes have a wet nose: the evolution of the use of guide dogs and establishing the seeing eye. Surv Ophthalmol. (2003) 48:452–8. doi: 10.1016/S0039-6257(03)00052-3

8. Cobb M, Branson N, McGreevy P, Lill A, Bennett P. The advent of canine performance science: offering a sustainable future for working dogs. Behav Process. (2015) 110:96–104. doi: 10.1016/j.beproc.2014.10.012

9. Cruse SD. Military working dogs: classification and treatment in the US Armed Forces. Animal. (2014) 21:249.

10. Levinson B. Pet-Oriented Child Psychotherapy . Springfield, IL: Charles C Thomas (1969).

11. Corson S, O'leary Corson E, Gwynne P. Pet-facilitated psychotherapy. ln: Anderson RS, editor. Petanimals and Society . London: Baillière Tindal (1975). p. 19–35.

12. Corson SA, Arnold LE, Gwynne PH, Corson EOL. Pet dogs as nonverbal communication links in hospital psychiatry. Compr Psychiatry. (1977) 18:61–72. doi: 10.1016/S0010-440X(77)80008-4

13. Fine AH. Handbook on Animal-Assisted Therapy: Foundations and Guidelines for Animal-Assisted Interventions . New York, NY: Academic Press (2019).

14. Engel GL. The clinical application of the biopsychosocial model. Am J Psychiatry. (1980) 137:535–44. doi: 10.1176/ajp.137.5.535

CrossRef Full Text | Google Scholar

15. Lehman BJ, David DM, Gruber JA. Rethinking the biopsychosocial model of health: understanding health as a dynamic system. Soc Person Psychol Compass. (2017) 11:e12328. doi: 10.1111/spc3.12328

16. Gee NR, Mueller MK. A systematic review of research on pet ownership and animal interactions among older adults. Anthrozoös. (2019) 32:183–207. doi: 10.1080/08927936.2019.1569903

17. Krause-Parello CA. Pet ownership and older women: the relationships among loneliness, pet attachment support, human social support, and depressed mood. Geriatr Nurs. (2012) 33:194–203. doi: 10.1016/j.gerinurse.2011.12.005

18. Kline JA, Fisher MA, Pettit KL, Linville CT, Beck AM. Controlled clinical trial of canine therapy versus usual care to reduce patient anxiety in the emergency department. PLoS ONE. (2019) 14:e0209232. doi: 10.1371/journal.pone.0209232

19. Kerns KA, Stuart-Parrigon KL, Coifman KG, Van Dulmen MH, Koehn A. Pet dogs: does their presence influence preadolescents' emotional responses to a social stressor? Soc Dev. (2018) 27:34–44. doi: 10.1111/sode.12246

20. Kertes DA, Liu J, Hall NJ, Hadad NA, Wynne CD, Bhatt SS. Effect of pet dogs on children's perceived stress and cortisol stress response. Soc Dev. (2016) 26:382–401. doi: 10.1111/sode.12203

21. Crossman MK, Kazdin AE, Matijczak A, Kitt ER, Santos LR. The influence of interactions with dogs on affect, anxiety, and arousal in children. J Clin Child Adolesc Psychol. (2020) 49:535–48. doi: 10.1080/15374416.2018.1520119

22. Crossman MK. Effects of interactions with animals on human psychological distress. J Clin Psychol. (2017) 73:761–84. doi: 10.1002/jclp.22410

23. Hunt MG, Chizkov RR. Are therapy dogs like Xanax? Does animal-assisted therapy impact processes relevant to cognitive behavioral psychotherapy? Anthrozoös. (2014) 27:457–69. doi: 10.2752/175303714X14023922797959

24. Bao KJ, Schreer G. Pets and happiness: examining the association between pet ownership and wellbeing. Anthrozoös. (2016) 29:283–96. doi: 10.1080/08927936.2016.1152721

25. Mcconnell AR, Brown CM, Shoda TM, Stayton LE, Martin CE. Friends with benefits: on the positive consequences of pet ownership. J Person Soc Psychol. (2011) 101:1239. doi: 10.1037/a0024506

26. Luhmann M, Kalitzki A. How animals contribute to subjective well-being: A comprehensive model of protective and risk factors. J Positive Psychol. (2018) 13:200–14. doi: 10.1080/17439760.2016.1257054

27. Taylor P, Funk C, Craighill P. “Are We Happy Yet?” . Washington, DC: Pew Research Center (2006).

28. Barcelos AM, Kargas N, Maltby J, Hall S, Mills DS. A framework for understanding how activities associated with dog ownership relate to human well-being. Sci Rep. (2020) 10:1–12. doi: 10.1038/s41598-020-68446-9

29. Cline KMC. Psychological effects of dog ownership: role strain, role enhancement, and depression. J Soc Psychol. (2010) 150:117–31. doi: 10.1080/00224540903368533

30. Wanser SH, Vitale KR, Thielke LE, Brubaker L, Udell MA. Spotlight on the psychological basis of childhood pet attachment and its implications. Psychol Res Behav Manag. (2019) 12:469. doi: 10.2147/PRBM.S158998

31. Barker SB, Wolen AR. The benefits of human–companion animal interaction: a review. J Vet Med Educ. (2008) 35:487–95. doi: 10.3138/jvme.35.4.487

32. Raina P, Waltner-Toews D, Bonnett B, Woodward C, Abernathy T. Influence of companion animals on the physical and psychological health of older people: An analysis of a one-year longitudinal study. J Am Geriatr Soc. (1999) 47:323–9. doi: 10.1111/j.1532-5415.1999.tb02996.x

33. Melson GF, Fine AH. Animals in the lives of children. In: Fine AH, editor. Handbook on Animal-Assisted Therapy. New York, NY: Elsevier (2015). p. 179–94.

34. Rodriguez KE, Greer J, Beck A, O'haire ME. The psychosocial effects of service dogs on individuals with physical disabilities: a systematic literature review. PLoS ONE. (2020) 15:e0243302. doi: 10.1371/journal.pone.0243302

35. Krause-Parello CA, Sarni S, Padden E. Military veterans and canine assistance for post-traumatic stress disorder: a narrative review of the literature. Nurse Educ Today. (2016) 47:43–50. doi: 10.1016/j.nedt.2016.04.020

36. Rodriguez KE, Lafollette MR, Hediger K, Ogata N, O'haire ME. Defining the PTSD service dog intervention: perceived importance, usage, and symptom specificity of psychiatric service dogs for military veterans. Front Psychol. (2020) 11:1638. doi: 10.3389/fpsyg.2020.01638

37. Silva K, Lima M, Santos-Magalhães A, Fafiães C, De Sousa L. Can dogs assist children with severe autism spectrum disorder in complying with challenging demands? An exploratory experiment with a live and a robotic dog. J Alternative Comp Med. (2018) 24:238–42. doi: 10.1089/acm.2017.0254

38. Wright H, Hall S, Hames A, Hardiman J, Mills R, Team PP, et al. Pet dogs improve family functioning and reduce anxiety in children with autism spectrum disorder. Anthrozoös. (2015) 28:611–24. doi: 10.1080/08927936.2015.1070003

39. Viau R, Arsenault-Lapierre G, Fecteau S, Champagne N, Walker C-D, Lupien S. Effect of service dogs on salivary cortisol secretion in autistic children. Psychoneuroendocrinology. (2010) 35:1187–93. doi: 10.1016/j.psyneuen.2010.02.004

40. Burrows KE, Adams CL, Spiers J. Sentinels of safety: Service dogs ensure safety and enhance freedom and well-being for families with autistic children. Qual Health Res. (2008) 18:1642–9. doi: 10.1177/1049732308327088

41. Carlisle GK. The social skills and attachment to dogs of children with autism spectrum disorder. J Autism Dev Disord. (2014) 45:1137–45.

PubMed Abstract | Google Scholar

42. Schuck SE, Emmerson NA, Fine AH, Lakes KD. Canine-assisted therapy for children with ADHD: preliminary findings from the positive assertive cooperative kids study. J Attention Disord. (2015) 19:125–37. doi: 10.1177/1087054713502080

43. Schuck SE, Johnson HL, Abdullah MM, Stehli A, Fine AH, Lakes KD. The role of animal assisted intervention on improving self-esteem in children with attention deficit/hyperactivity disorder. Front Pediatr. (2018) 6:300. doi: 10.3389/fped.2018.00300

44. Grandin T, Fine AH, Bowers CM. The use of therapy animals with individuals with autism spectrum disorders. In: Fine AH, edior. Handbook on Animal-Assisted Therapy . 3rd ed. New York, NY: Elsevier. (2010). p. 247–64.

45. Hall SS, Gee NR, Mills DS. Children reading to dogs: a systematic review of the literature. PLoS ONE. (2016) 11. doi: 10.1371/journal.pone.0149759

46. Gee NR, Fine AH, Schuck S. Animals in educational settings: Research and practice. In: Handbook on Animal-Assisted Therapy . New York, NY: Elsevier (2015). p. 195–210.

47. Hediger K, Turner D. Can dogs enhance children's attention performance? A randomized controlled crossover trial. Hum Anim Interact Bull. (2014) 2:21–39.

48. Ward-Griffin E, Klaiber P, Collins HK, Owens RL, Coren S, Chen FS. Petting away pre-exam stress: the effect of therapy dog sessions on student well-being. Stress Health. (2018) 34:468–73. doi: 10.1002/smi.2804

49. Wood E, Ohlsen S, Thompson J, Hulin J, Knowles L. The feasibility of brief dog-assisted therapy on university students stress levels: the PAWS study. J Mental Health. (2017) 27:263–8. doi: 10.1080/09638237.2017.1385737

50. Binfet J-T, Passmore H-A, Cebry A, Struik K, Mckay C. Reducing university students' stress through a drop-in canine-therapy program. J Mental Health. (2018) 27:197–204. doi: 10.1080/09638237.2017.1417551

51. Pendry P, Carr AM, Gee NR, Vandagriff JL. Randomized trial examining effects of animal assisted intervention and stress related symptoms on college students' learning and study skills. Int J Environ Res Public Health. (2020) 17:1909. doi: 10.3390/ijerph17061909

52. Trammell JP. The effect of therapy dogs on exam stress and memory. Anthrozoös. (2017) 30:607–21. doi: 10.1080/08927936.2017.1370244

53. Trammell JP. Therapy dogs improve student affect but not memory. Anthrozoös. (2019) 32:691–9. doi: 10.1080/08927936.2019.1645514

54. Rodriguez KE, Guérin NA, Gabriels RL, Serpell JA, Schreiner PJ, O'haire ME. The state of assessment in human-animal interaction research. Hum Anim Interact Bull. (2018) 6:63−81.

55. Beetz A, Uvnäs-Moberg K, Julius H, Kotrschal K. Psychosocial and psychophysiological effects of human-animal interactions: the possible role of oxytocin. Front Psychol. (2012) 3:234. doi: 10.3389/fpsyg.2012.00234

56. Odendaal J, Meintjes R. Neurophysiological correlates of affiliative behaviour between humans and dogs. Vet J. (2003) 165:296–301. doi: 10.1016/S1090-0233(02)00237-X

57. Pendry P, Vandagriff JL. Salivary Studies of the Social Neuroscience of Human–Animal Interaction. In: Granger D, Taylor M, editors. Salivary Bioscience. Springer (2020). p. 555–81.

58. Barker SB, Knisely JS, Mccain NL, Best AM. Measuring stress and immune response in healthcare professionals following interaction with a therapy dog: a pilot study. Psychol Rep. (2005) 96:713–29. doi: 10.2466/pr0.96.3.713-729

59. Rodriguez KE, Bryce CI, Granger DA, O'haire ME. The effect of a service dog on salivary cortisol awakening response in a military population with posttraumatic stress disorder (PTSD). Psychoneuroendocrinology. (2018) 98:202–10. doi: 10.1016/j.psyneuen.2018.04.026

60. Polheber J, Matchock R. (2013). The presence of a dog attenuates cortisol and heart rate in the Trier Social Stress Test compared to human friends. J Behav Med. 37:1–8. doi: 10.1007/s10865-013-9546-1

61. Meints K, Brelsford V, Dimolareva M, Gee N. Dog-assisted interventions with children in mainstream and special educational needs schools – what works? In: Paper presented at the International Society for Anthrozoology Annual Conference . Liverpool (2020).

62. Coakley AB, Annese CD, Empoliti JH, Flanagan JM. The experience of animal assisted therapy on patients in an acute care setting. Clin Nurs Res. (2020). doi: 10.1177/1054773820977198. [Epub ahead of print].

63. Clark S, Martin F, Mcgowan RT, Smidt J, Anderson R, Wang L, et al. The impact of a 20-minute animal-assisted activity session on the physiological and emotional states in patients with fibromyalgia. In: Mayo Clinic Proceedings . Elsevier. (2020). p. 2442−61. doi: 10.1016/mayocp.2020.04.037

64. Lass-Hennemann J, Schäfer SK, Römer S, Holz E, Streb M, Michael T. Therapy dogs as a crisis intervention after traumatic events? An experimental study. Front Psychol. (2018) 9:1627. doi: 10.3389/fpsyg.2018.01627

65. Bleiberg J, Prout M, Debiak D, Lefkowitz C, Paharia I. Animal-assisted prolonged exposure: a treatment for survivors of sexual assault suffering posttraumatic stress disorder. Soc Anim. (2005) 13:275–96. doi: 10.1163/156853005774653654

66. O'haire ME, Philip Tedeschi M, Jenkins MA, Braden SR, Rodriguez KE. The impact of human-animal interaction in trauma recovery. New Direct Hum Anim Bond . (2019) 15:15–53. Available online at: https://library.oapen.org/bitstream/handle/20.500.12657/24992/1005110.pdf?sequence=1#page=30

67. Beetz A, Julius H, Turner D, Kotrschal K. Effects of social support by a dog on stress modulation in male children with insecure attachment. Front Psychol. (2012) 3:352. doi: 10.3389/fpsyg.2012.00352

68. Cardoso C, Kingdon D, Ellenbogen MA. A meta-analytic review of the impact of intranasal oxytocin administration on cortisol concentrations during laboratory tasks: moderation by method and mental health. Psychoneuroendocrinology. (2014) 49:161–70. doi: 10.1016/j.psyneuen.2014.07.014

69. Meyer-Lindenberg A, Domes G, Kirsch P, Heinrichs M. Oxytocin and vasopressin in the human brain: social neuropeptides for translational medicine. Nat Rev Neurosci. (2011) 12:524–38. doi: 10.1038/nrn3044

70. Marshall-Pescini S, Schaebs FS, Gaugg A, Meinert A, Deschner T, Range F. The role of oxytocin in the dog–owner relationship. Animals. (2019) 9:792. doi: 10.3390/ani9100792

71. Handlin L, Nilsson A, Ejdebäck M, Hydbring-Sandberg E, Uvnäs-Moberg K. Associations between the psychological characteristics of the human–dog relationship and oxytocin and cortisol levels. Anthrozoos. (2012) 25:215–28. doi: 10.2752/175303712X13316289505468

72. Maclean EL, Gesquiere LR, Gee NR, Levy K, Martin WL, Carter CS. Effects of affiliative human–animal interaction on dog salivary and plasma oxytocin and vasopressin. Front Psychol. (2017) 8:1606. doi: 10.3389/fpsyg.2017.01606

73. Petersson M, Uvnäs-Moberg K, Nilsson A, Gustafson L-L, Hydbring-Sandberg E, Handlin L. Oxytocin and cortisol levels in dog owners and their dogs are associated with behavioral patterns: an exploratory study. Front Psychol. (2017) 8:1796. doi: 10.3389/fpsyg.2017.01796

74. Miller SC, Kennedy C, Devoe D, Hickey M, Nelson T, Kogan L. An examination of changes in oxytocin levels in men and women before and after interaction with a bonded dog. Anthrozoös. (2009) 22:31–42. doi: 10.2752/175303708X390455

75. Powell L, Guastella AJ, Mcgreevy P, Bauman A, Edwards KM, Stamatakis E. The physiological function of oxytocin in humans and its acute response to human-dog interactions: a review of the literature. J Vet Behav. (2019) 30:25–32. doi: 10.1016/j.jveb.2018.10.008

76. Garcia MCS, Schorr AR, Arnold W, Fei N, Gilbert JA. Pets as a novel microbiome-based therapy. In: Pastorinho M, Sousa A, editors. Pets as Sentinels, Forecasters and Promoters of Human Health. Springer (2020). p. 245–67. doi: 10.1007/978-3-030-30734-9

77. Hölscher B, Frye C, Wichmann HE, Heinrich J. Exposure to pets and allergies in children. Pediatr Aller Immunol. (2002) 13:334–41. doi: 10.1034/j.1399-3038.2002.02063.x

78. Bufford J, Reardon C, Li Z, Roberg K, Dasilva D, Eggleston P, et al. Effects of dog ownership in early childhood on immune development and atopic diseases. Clin Exp Allergy. (2008) 38:1635–43. doi: 10.1111/j.1365-2222.2008.03018.x

79. Friedmann E. The animal-human bond: health and wellness. In: Fine AH, editor. Animal-Assisted Therapy: Theoretical Foundations and Guidelines for Practice . New York, NY: Academic Press. (2019). p. 80–93.

80. Kramer CK, Mehmood S, Suen RS. Dog ownership and survival: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes. (2019) 12:e005554. doi: 10.1161/CIRCOUTCOMES.119.005554

81. Beck AM, Katcher AH. Between Pets and People: The Importance of Animal Companionship . West Lafayette, IN: Purdue University Press (1996).

82. Kurdek LA. Pet dogs as attachment figures. J Soc Person Relat. (2008) 25:247–66. doi: 10.1177/0265407507087958

83. Powell L, Edwards KM, Mcgreevy P, Bauman A, Podberscek A, Neilly B, et al. Companion dog acquisition and mental well-being: a community-based three-arm controlled study. BMC Public Health. (2019) 19:1428. doi: 10.1186/s12889-019-7770-5

84. Oliva JL, Johnston KL. Puppy love in the time of Corona: dog ownership protects against loneliness for those living alone during the COVID-19 lockdown. Int J Soc Psychiatry. (2020). doi: 10.1177/0020764020944195. [Epub ahead of print].

85. Stanley IH, Conwell Y, Bowen C, Van Orden KA. Pet ownership may attenuate loneliness among older adult primary care patients who live alone. Aging Mental Health. (2014) 18:394–9. doi: 10.1080/13607863.2013.837147

86. Krause-Parello CA, Gulick EE, Basin B. Loneliness, depression, and physical activity in older adults: the therapeutic role of human–animal interactions. Anthrozoös. (2019) 32:239–54. doi: 10.1080/08927936.2019.1569906

87. Gilbey A, Tani K. Companion animals and loneliness: a systematic review of quantitative studies. Anthrozoos. (2015) 28:181–97. doi: 10.2752/089279315X14219211661615

88. Mcnicholas J, Collis GM. Dogs as catalysts for social interaction: Robustness of the effect. Br J Psychol. (2000) 91:61–70. doi: 10.1348/000712600161673

89. Mader B, Hart LA, Bergin B. Social acknowledgments for children with disabilities: effects of service dogs. Child Dev. (1989) 60:1529–34. doi: 10.2307/1130941

90. Higgins JW, Temple V, Murray H, Kumm E, Rhodes R. Walking sole mates: dogs motivating, enabling and supporting guardians' physical activity. Anthrozoös. (2013) 26:237–52. doi: 10.2752/175303713X13636846944286

91. Lang RE, Hornburg SP. What is social capital and why is it important to public policy? Housing Policy Debate. (1998) 9:1–16. doi: 10.1080/10511482.1998.9521284

92. Wood L, Giles-Corti B, Bulsara M. The pet connection: pets as a conduit for social capital? Soc Sci Med. (2005) 61:1159–73. doi: 10.1016/j.socscimed.2005.01.017

93. Purewal R, Christley R, Kordas K, Joinson C, Meints K, Gee N, et al. Companion animals and child/adolescent development: a systematic review of the evidence. Int J Environ Res Public Health. (2017) 14:234. doi: 10.3390/ijerph14030234

94. Jalongo MR. An attachment perspective on the child-dog bond: Interdisciplinary and international research findings. Children Dogs Educ. (2018) 21–41. doi: 10.1007/978-3-319-77845-7_2

95. Melson GF. Companion animals and the development of children: implications of the biophilia hypothesis. In: Fine AH, editor. Handbook on Animal-Assisted Therapy . New York, NY: Elsevier (2006). p. 375–83. doi: 10.1016/B978-012369484-3/50019-0

96. Mueller MK. Is human-animal interaction (HAI) linked to positive youth development? Init Answers Appl Dev Sci. (2014) 18:5–16. doi: 10.1080/10888691.2014.864205

97. Vidović VV, Štetić VV, Bratko D. Pet ownership, type of pet and socio-emotional development of school children. Anthrozoös. (1999) 12:211–7. doi: 10.2752/089279399787000129

98. Wenden EJ, Lester L, Zubrick SR, Ng M, Christian HE. The relationship between dog ownership, dog play, family dog walking, and pre-schooler social–emotional development: findings from the PLAYCE observational study. Pediatr Res. 1–7. doi: 10.1038/s41390-020-1007-2. [Epub ahead of print].

99. Prothmann A, Ettrich C, Prothmann S. Preference for, and responsiveness to, people, dogs and objects in children with autism. Anthrozoös. (2009) 22:161–71. doi: 10.2752/175303709X434185

100. Solomon O. What a dog can do: Children with autism and therapy dogs in social interaction. Ethos. (2010) 38:143–66. doi: 10.1111/j.1548-1352.2010.01085.x

101. Valiyamattam GJ, Katti H, Chaganti VK, O'haire ME, Sachdeva V. Do animals engage greater social attention in autism? An eye tracking analysis. Front Psychol. (2020) 11:727. doi: 10.3389/fpsyg.2020.00727

102. Martin F, Farnum J. Animal-assisted therapy for children with pervasive developmental disorders. West J Nurs Res. (2002) 24:657–70. doi: 10.1177/019394502320555403

103. Silva K, Correia R, Lima M, Magalhães A, De Sousa L. Can dogs prime autistic children for therapy? Evidence from a single case study. J Alternat Comp Med. (2011) 17:1–5. doi: 10.1089/acm.2010.0436

104. Carlisle GK. Pet dog ownership decisions for parents of children with autism spectrum disorder. J Pediatr Nurs. (2014) 29:114. doi: 10.1016/j.pedn.2013.09.005

105. Hall SS, Wright HF, Hames A, Mills DS. The long-term benefits of dog ownership in families with children with autism. J Vet Behav. (2016) 13:46–54. doi: 10.1016/j.jveb.2016.04.003

106. Beck AT, Steer RA, Brown G. Beck depression inventory–II. Psychol Assess. (1996) 36:221–37. doi: 10.1037/t00742-000

107. Russell DW. UCLA Loneliness Scale (Version 3): reliability, validity, and factor structure. J Person Assess. (1996) 66:20–40. doi: 10.1207/s15327752jpa6601_2

108. Van Der Oord S, Van Der Meulen E, Prins PJ, Oosterlaan J, Buitelaar J, Emmelkamp PM. A psychometric evaluation of the social skills rating system in children with attention deficit hyperactivity disorder. Behav Res Therapy. (2005) 43:733–46. doi: 10.1016/j.brat.2004.06.004

109. Herzog H. The impact of pets on human health and psychological well-being: fact, fiction, or hypothesis? Curr Direct Psychol Sci. (2011) 20:236–9. doi: 10.1177/0963721411415220

110. Rodriguez KE, Herzog H, Gee N. Variability in human-animal interaction research. Front Vet Sci. (2021) 7:619600. doi: 10.3389/fvets.2020.619600

111. Saleh US. Theory guided practice in nursing. J Nurs Res Pract. (2018) 2:18.

Keywords: dog, human health, human-animal interaction, biopsychosocial, canine, mental health

Citation: Gee NR, Rodriguez KE, Fine AH and Trammell JP (2021) Dogs Supporting Human Health and Well-Being: A Biopsychosocial Approach. Front. Vet. Sci. 8:630465. doi: 10.3389/fvets.2021.630465

Received: 17 November 2020; Accepted: 25 February 2021; Published: 30 March 2021.

Reviewed by:

Copyright © 2021 Gee, Rodriguez, Fine and Trammell. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Nancy R. Gee, nancy.gee@vcuhealth.org

This article is part of the Research Topic

Our Canine Connection: The History, Benefits and Future of Human-Dog Interactions

Science explores the origins of the friendship between dogs and humans

Recent studies confirm dogs’ ability to understand us, their natural talent for empathizing with other species and the pleasure we get from sharing our lives with them.

It is often said that dogs are man’s best friend, but less often that they are the oldest. Dogs were the first domesticated animal in history. The two species stitched their evolutionary destinies together some 15,000 years ago, establishing a symbiotic relationship with few analogues in the animal world. It’s a rarity. Decades ago, archaeologists and zoologists posited that this relationship was born of utility but that, over the years, a fondness and understanding emerged between the two species that science is now trying to measure. In recent years, several studies have analyzed how this joint evolution has affected dogs and humans. In the past 20 years, the scientific literature on this subject has proliferated. And so has the conventional one. It is estimated that there are over 70,000 books about dogs on Amazon : a further sign that this prehistoric friendship is coming to the present in top form.

Onyoo Yoo has a beautiful four-year-old poodle named Aroma; they call her Aro at home. There were others before: Yoo has spent her whole life surrounded by dogs and knows from personal experience that these animals can bring joy and comfort, although she doesn’t quite understand the mechanisms that make this possible. Last year, Yoo took her dog to work to find out. She asked 30 volunteers to pet Aro, give her treats, walk her and play with her. Meanwhile, Yoo, who is a researcher at Konkuk University in South Korea, analyzed their brain activity.

“Our research found that participants’ alpha band brain waves [related to relaxation] increased while playing and walking with my dog. While beta band brain waves [associated with concentration] did so while grooming, massaging or playing with her,” Yoo explains. Recently published in the scientific journal PLOS One , the study confirms something many people feel : spending time with dogs is enormously pleasurable. But it does so in detail, “providing valuable information to elucidate the therapeutic effects and underlying mechanisms of animal-assisted interventions,” Yoo explains.

Pet ownership is known to help reduce stress levels, promote positive emotions and reduce the risk of cardiovascular disease. “However, research on the brain activity produced by human-animal interaction is incipient and insufficient,” says Yoo. This may be because, in order to understand it, one needs not only neurology and psychology but also paleobiology, in addition to taking a look back.

Initiating a friendship is not always easy, and the one forged between people and dogs did not start by petting a wolf or throwing a ball at it. Domestication was multifactorial and happened in fits and starts. An ambitious study published in Science in 2020 attempted to trace that process by sequencing 27 ancient dog genomes . In analyzing them, the authors discovered that dogs probably arose from a now-extinct population of wolves. They also distinguished at least five different dog populations, drawing a complex ancestral history. Different types of dogs expanded with the other human groups, linking their destinies (and eventual disappearance) to the survival of the clan with which they were associated.

Aritza Villaluenga, a researcher at the University of the Basque Country UPV/EHU (Spain) and a co-author of the study, points out that the first (albeit disputed) evidence of coexistence between humans and wolves dates back to 25,000 years ago: “It was probably not a conscious domestication, they did not know what they were doing, they did not conceive of what the result was going to be. They simply had animals to help them hunt.” We have to jump forward 10,000 years to when dogs first appeared in a sustained way. “And here, yes, we can talk about dogs because genetically they’re different from the wolves living in the same area at the same time. There [were] physical and genetic changes,” Villaluenga explains.

Hunting allies

At that time, their coexistence was symbiotic. “The partnership suited [both] dogs and humans. The dogs pushed herds of animals to where the human hunters were hiding,” the expert notes. The former were much better at running and the latter at strategizing. They formed a good team when hunting and, once the game was caught, they both shared the spoils. Thus, from the beginning of the relationship, it was very important for both species to understand each other, to be able to read each other.

Researchers at Stockholm University (Sweden) conducted an experiment with wolf cubs and found that some specimens can understand human cues and comprehend their playful intentionality. They proved this with something as trivial as throwing them a ball and asking them to bring it back. In everyday life, this action seems simple because many people do it with their pets daily. But that task involves great cognitive complexity, and in a few seconds, it demonstrates the ability of two species to understand each other that has been forged over millennia. The study suggested that this ability, present in some very sociable specimens, may have led to their domestication. Associating with humans was an evolutionary success in all respects. Today, it is estimated that for every wolf, there are 3,000 dogs.

Some 5,000 generations after that prehistoric union, today’s dogs can understand many more human commands, gestures and words. Mariana Boros, a neuroethologist at Loránd University in Budapest (Hungary), knows that well. She just published a study that analyzes how dogs can understand words . “The most important ability this animal has is the ability to understand human communication. They are exceptional,” the expert observes in a video call.

Boros and her team wanted to test whether this understanding was due to vocalization rather than context. So, they locked a dog in a room, announced that they were going to give him one object, say a ball, and then offered him another, for example, a stick. “We thought that if the dog understood what the word meant, it would have an expectation of what it would see next. And the violation of that expectation would be visible on the EEG,” Boros explains. And, indeed, it was. With this data, the team can be sure that dogs understand the word’s meaning. “In fact, the understanding mechanisms are very similar to what we see in humans,” Boros adds.

Love beyond understanding

Most scientific literature concludes that dogs have a special bond with humans for this reason. They understand us and communicate with us better than any other animal. But psychologist Clive Wynne of the University of Arizona disagrees. In his book Dog is Love , he argues that dogs have a unique capacity for interspecies love. If you raise a dog with sheep, goats or cats (or even tigers or lions), it will end up joining them and getting attached to them, he explains, using examples. Something similar would have happened with humans. Wyne’s hypothesis is backed up by science. In 2015, Japanese scientists showed that the more people looked into their dogs’ eyes, the more they both increased their production of oxytocin, the key chemical ingredient of affection. It’s not that they understand the humans with whom they live; it is that they love them .

In any case, understanding is not the only aspect in which dogs have evolved to adapt to our tastes. According to several studies, they have also become more adorable and expressive. Charles Darwin was the first to realize that domestic animals — such as cats, dogs and rabbits — share certain physical traits. They tend to have droopier ears and curlier tails than their wild ancestors. Their teeth are smaller, and they have white patches on their fur. This phenomenon is known as domestication syndrome.

The most eloquent example of this process occurred at a Soviet fox farm in the 1950s. Geneticist Dmitri K. Belyaev wanted to create a domestic fox population by selecting and crossing the tamest specimens. The results were analyzed in a scientific study in 2009 . By the fourth generation, the foxes were licking scientists and greeting them with tail wags. Their offspring were even more domesticated and could understand human signals and respond to gestures or looks. “They not only developed internal traits such as acceptance of human closeness. Physically, they became more pup-like, cuter. They changed to be more adorable to the human eye and presumably that same thing happened to dogs,” Boros explains.

The difference is that, with the foxes, this happened artificially and forcibly in just 50 years, while the domestication of the wolf into a dog was natural and presumably took much longer. This process was not born of man’s whim, as Villaluenga explained. Some Stone Age wolves showed a natural inclination to befriend the strange apes that spread throughout the world. They understood each other not only when hunting, but also when playing and being affectionate with each other. When they looked at each other, they both felt strangely good. These wolves got closer and closer to the humans and mingled with other wolves that were also hanging around human settlements. They decided to stay close, and it turned out to be for the best. According to this interpretation, shared by many specialists, the dog was not domesticated, but rather some wolves became self-domesticated and ended up becoming dogs. They chose us, at least as much as we chose them.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition

More information

Do dogs dream? What about?

Is it healthy for dogs to eat cereals?

• Francés online
• Inglés online
• Italiano online
• Alemán online
• Crucigramas & Juegos

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

•  We're Hiring!
•  Help Center
• Most Cited Papers
• Most Downloaded Papers
• Newest Papers
• Save to Library
• Last »
• Dog Behavior Follow Following
• Dog Cognition Follow Following
• Human-Dog Relations Follow Following
• Behavior of Dogs, Evolution of Dogs, Domestication Processes Follow Following
• Cats Follow Following
• Archaeological dogs Follow Following
• PET Follow Following
• Canine ethology, behavior, and training Follow Following
• Guilty Dog Follow Following
• Human-Animal Relations Follow Following

Enter the email address you signed up with and we'll email you a reset link.

• Academia.edu Publishing
•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

Read our Dog Testing FAQ

Our new Dog FAQ page provides answers to the most frequently asked questions about dogs in research.

Dogs ( Canis familiaris ) belong to the family Canidae and are thought to be one of the first domesticated animals. They have been used in research for more than a century, however, they are currently very rarely used in animal research in Great Britain, only being used in 0.24% of experimental procedures in 2019 (latest published figures).

They are medium-sized mammals that can grow from 15 to 100 cm and weigh from 1.5 to 75 kg, depending on the type of dog. Dogs are carnivores but can thrive on a well-designed suitably processed omnivorous diet in the domestic situation.

Why are dogs used in research?

In the UK, dogs are primarily used to find out how new drugs act within a whole, living body and whether new medicines are safe enough to test in humans. They predict this safety very well, with  up to 96% accuracy .

This is done to satisfy safety regulations which came about after the drug Thalidomide maimed and killed children while they were still in the womb. It is known as toxicology testing, but normally seeks to confirm the absence of toxic effects.

The tests can tell us lots of sorts of information all at once, like the safety if a drug across lots of different internal organs, how the drug travels around the body and other information that helps us to design much safer human trials.

Dogs are also used to test the safety and efficacy of veterinary medicines, and also in nutrition studies to ensure that pet dogs eat healthily, particularly when they are prescribed specialist diets by their vets.

Although animal and nonanimal methods are used alongside each other, there are currently no alternatives to using dogs. They nevertheless have special protections under UK law. For instance, they cannot be used if another animal species could be used.

There is a project that hopes to create a  ‘virtual dog’  that could significantly reduce the number of dogs needed by using computers to mine historical dog data. It is being run by the UK’s national centre for  developing animal replacements, the NC3Rs , but is of international interest.

What types of research are dogs used in?

The physiological similarities between humans and dogs mean that they are useful in various types of research. Their genome has been sequenced and because of our genetic similarities, they are often used in genetic studies.

Dogs are primarily used in regulatory research, also known as toxicology or safety testing. This type of research is required by law to test the safety and effectiveness of potential new medicines and medical devices before they are given to human volunteers during clinical trials. Dogs are also used to test the safety and efficacy of veterinary medicines, and also in nutrition studies to ensure that pet dogs eat healthily, particularly when they are prescribed specialist diets by their vets.

A smaller number of dogs are also used in translational research (also called applied research) to help us learn about human and animal diseases so that we can develop treatments. Examples of translational diseases can be found below.

Dogs are also used to study Duchenne muscular dystrophy (DMD), which is the most common type of muscular dystrophy. It is another condition that can affect both humans and dogs. Because dogs can naturally have this condition, they can be studied to show how the condition progresses. This very useful model for DMD has helped scientists work on better genetic tests and treatments for the condition.  

An early use of dogs in research was in the search for a treatment for diabetes, which resulted in the discovery of insulin. This discovery in the 1920s, which won researchers a Nobel prize, now allows people with diabetes to live long lives. In the past, people with diabetes would die soon after developing the condition.

How are the dogs looked after?

The use of animals in research is highly regulated, an important part of that regulation is ensuring the animals are housed and cared for correctly. Laboratory dogs are housed in enclosures that can isolate individual dogs for treatment but usually opened up for dogs to interact. Dogs’ need to socialise is well considered, so the dogs are housed in small groups most of the time. The facilities usually also have space to run around for exercise and you can usually find dogs interacting with each other, environmental enrichments, and the animal technicians.

https://www.nc3rs.org.uk/3rs-resources/housing-and-husbandry/housing-and-husbandry-dogs

You can see this in this film about dogs in research.

https://www.nationalgeographic.com/animals/mammals/d/domestic-dog/

https://www.britannica.com/animal/dog/Breed-specific-behaviour#ref15478

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5070630/#:~:text=Dogs%20have%20approximately%20the%20same,similar%20to%20human%20than%20mouse

http://www.animalresearch.info/en/designing-research/research-animals/dog/

Dogs in drug safety prediction:  https://pubmed.ncbi.nlm.nih.gov/28893587/

Virtual dog:  https://nc3rs.org.uk/news/ps16m-awarded-develop-virtual-second-species

Featured news

UK government doubles funding for the 3Rs

Letter from Lord Sharpe of Epsom responding to the ASC forced swim test report

UK animal rights extremists cleared of ‘psychological warfare’ on businessman

Subscribe to our newsletter.

Get the latest articles and news from Understanding Animal Research in your email inbox every month. For more information, please see our  privacy policy .

Premium Content

Dog DNA tests are on the rise—but are they reliable?

Several companies now offer services to sequence Fido’s genetic information—but how do they really work?

Could your dog be predisposed to a fatal disease? Is your new shelter pup part beagle or boxer? Many pet owners seek answers to these questions, and as a result, direct-to-consumer dog DNA testing is booming.

But how reliable are dog DNA tests—and are the results worth your time? As it turns out, swabbing your dog’s slobber and mailing it to a testing service is the easy part, but just how (or whether) to act on that information presents some ruff choices for scientists and pet owners alike.

Domestication and dog DNA

Some scientists speculate that dogs separated off from wolves about 23,000 years ago while others claim it happened about 10,000 years later. Either way, humans have left an indelible impression on canines: Careful dog breeding—often tied to desired physical or behavioral characteristics—has produced nearly 400 modern breeds worldwide.

Scientists sequenced a full dog genome for the first time in 2004. Since then, we’ve learned much more about canines’ genetic predisposition to a variety of conditions such as kidney cancer, retinal atrophy, and hip dysplasia.

In one massive 2023 study of more than 1 million dogs, researchers screened for 250 genetic variants associated with diseases like bald thigh syndrome, a hair loss disorder that primarily strikes hounds, and cone-rod dystrophy, an eye disorder that can lead to blindness among pit bulls. The researchers found that 57 percent of dogs carry at least one studied disease variant, and that the less genetic variability a dog has, the more disease markers in its DNA—which increases the chance of health issues.

Most dogs are highly inbred, research suggests: A 2021 study across 227 breeds found high levels of inbreeding across the board, and suggests that less inbred dogs are healthier than their   more inbred counterparts. Many breeders and veterinarians use DNA tests to detect potential problems in purebred dogs, and many vets recommend that owners of both mixed-breed and purebred dogs with shepherd ancestry get DNA testing to look for a potential genetic anomaly that can cause these dogs to experience neurological symptoms when they take some common veterinary drugs like ivermectin. But ultimately, the decision to test your dog for potential health issues is up to personal preference.

How reliable are dog DNA tests?

Human meddling with dog DNA has long been the driving factor behind dogs’ breed diversity—or lack thereof. But a dog’s DNA can also be used to confirm their lineage or identify their breed, a boon for pet owners on the lookout for breed-specific health or behavioral challenges or those looking to confirm their dog really has the heritage claimed by a breeder or seller. During DNA analysis, labs sequence the dog’s DNA and look for similarities with a dataset of identified dog breeds.

But breed identification isn’t as simple as it might seem. In a study published in the Journal of the American Veterinary Medical Association last month, scientists looked into the accuracy of breed prediction in commercially available DNA tests that required a photo of the dog in addition to its DNA sample. The results were mixed, says Casey Greene, a professor of biomedical informatics at the University of Colorado who co-authored the study.

“Most tests could accurately distinguish the breed of purebred dogs,” says Greene. But the analysis suggested that some testing companies might rely on the photo more than the dog’s actual genetics—and revealed big differences between companies’ business practices and the genetic datasets they use to determine dog breeds.

You May Also Like

What’s the real difference between a mutt and a designer dog?

Why your dog helps you relax more than your friends do

Meet Wicked—the very good girl who was just named America's best working dog

The researchers submitted photos and DNA of 12 purebred dogs to a total of six commercial canine ancestry identification services. Since each pup was purebred and possessed extensive American Kennel Club paperwork, the researchers knew their breed conclusively—but in some cases they provided a photo of a different dog to see if the photo influenced the DNA results. One of the companies misidentified a purebred Chinese crested dog—almost entirely hairless—as a long-haired Brittany spaniel, seemingly based on the photo alone. The other five did identify the registered breed correctly, but often gave different predictions for other “ancestor” breeds in dogs whose DNA suggested mixed breeding in prior generations. The researchers concluded that veterinarians and pet owners alike should “approach [direct-to-consumer] tests with caution” given the lack of industry standardization and at least one company’s reliance on photographs instead of DNA analysis.

What’s in a breed?

“Breed is a surprisingly complicated question,” says Halie Rando , an assistant professor of computer science at Smith College who led the research. A dog’s genetics may point to one breed, she says. But widely accepted breed definitions were defined in a time before DNA analysis, and Rando says that genetic testing can sometimes clash with pet owners’ preconceived notions about their dogs.

Even experienced humans, it turns out, are terrible at identifying breed by sight: In a recent study of 459 shelter dogs at two humane societies in Arizona and California, DNA analysis pinpointed found 125 distinct breeds, including five percent that were purebred. Nonetheless, neither the scientists nor the experienced shelter workers were able to reliably identify mixed-breed dogs, which made up nearly 90 percent of the canine cohort.

Mixed breeds can prove tricky even with DNA data—and since genetic testing relies on information about the genes of dogs with identifiable breeds, a DNA test is only as good as its genetic dataset.

“As a consumer, you might value a company that is more transparent and has a diverse [DNA] panel,” says Greene. He encourages consumers to do their research before submitting a sample—and double-check that the panel used by the testing company includes the breeds you suspect might be in the mix.

Even if you do get accurate information about your dog’s breed, it might not be as linked to its behavior as you might think. A 2022 genetic analysis of more than 2,000 purebred and mixed-breed dogs found behavior was linked more closely to individual dogs than breeds, concluding that “dog breed is generally a poor predictor of individual behavior.”

Given the rise in targeting certain breeds perceived as dangerous, there’s a chance that the growing direct-to-consumer testing market could perpetuate breed stereotypes. In their analysis, Rando and Greene anticipate the “social and economic consequences” of direct-to-consumer testing, including the potential for such tests to be used in decision-making that could affect people’s housing, insurance coverage, and even their ability to live in certain locations with dogs whose breeds are forbidden under local law or flagged by insurers.

Lab’s best friend

Despite these concerns, though, dog DNA seems headed for a golden age—and the insights revealed through further study of Fido’s genome have already reached far beyond the doghouse. Domesticated dogs have emerged as surprising superstars in medical research that benefits humans. For example, in 1999 researchers managed to isolate a genetic mutation that causes narcolepsy in some breeds. That breakthrough then led to the discovery of a similar mutation in people—and is driving better treatment for the disease in both humans and dogs alike.

“It’s fun to learn about your dog,” says Rando. But these days, she adds, that’s just the beginning. With implications ranging from entertaining to consequential, there’s no telling what dog DNA will continue to unleash.

Related Topics

• DOMESTIC DOGS
• ANIMAL REPRODUCTION
• ANIMAL DISEASES

Dogs will eat almost anything—cats, not so much. Here’s why.

Who was the Coast Salish woolly dog? DNA testing provides some clues.

Meet ‘Dogxim,’ the world’s first known dog-fox hybrid—and a genetic oddity

Radioactive dogs? What we can learn from Chernobyl's strays

Inside the growing U.S. movement to breed healthier, friendlier dogs

• History & Culture
• Environment
• Paid Content

History & Culture

• History Magazine
• Mind, Body, Wonder
• Terms of Use
• Privacy Policy
• Your US State Privacy Rights
• Children's Online Privacy Policy
• Interest-Based Ads
• About Nielsen Measurement
• Do Not Sell or Share My Personal Information
• Nat Geo Home
• Attend a Live Event
• Book a Trip
• Inspire Your Kids
• Shop Nat Geo
• Visit the D.C. Museum
• Learn About Our Impact
• Support Our Mission
• Advertise With Us
• Customer Service
• Renew Subscription
• Manage Your Subscription
• Work at Nat Geo
• Sign Up for Our Newsletters
• Contribute to Protect the Planet

Copyright © 1996-2015 National Geographic Society Copyright © 2015-2024 National Geographic Partners, LLC. All rights reserved

Yes, dogs can understand, link objects to words, researchers say

Like humans, dogs have the capacity to link words to mental images or ideas in their minds, researchers have recently found.

Researchers in Hungary and Norway made the discovery while researching brain activity in dogs. 

They had dog owners show their pets toys while playing recordings referencing each toy. Sometimes the toys they held matched the words the dogs heard. Other times, the toys were different from the words spoken.

The dogs’ brains responded differently when the owner said a toy name but held up a different object. 

“In the case of the different toys, the response wave is bigger than the response wave for the matching object,” said Marianna Boros, a co-lead author on the study and postdoctoral researcher at Eötvös Loránd University’s Neuroethology of Communication Lab in Hungary.

“This is something that has been also observed in humans. … This is the first time we were able to demonstrate something similar in a nonhuman mammal,” she said.

She added that the team invited dogs whose owners said the animals know at least three object names. The study’s sample also included a dog who knew over 230 object names.

“If a dog will learn an object name … it links a word to a representation in its mind and not just contextually associate and try to figure out what is expected from them,” said co-lead author Lilla Magyari, a researcher at Eötvös Loránd University’s Neuroethology of Communication Lab in Hungary and an associate professor at Stavanger University in Norway.

The findings are set to be published in the peer-reviewed journal Current Biology .

Dogs and their understanding of the human language

The researchers at the Neuroethology of Communication Lab took on the study because they are interested in how the brains of different mammal species process speech and voices, as well as the social cognition of different mammals.

Dogs are a “very special” species because they've lived in close proximity to humans for at least 18,000 to 30,000 years, said Boros. 

“They not only live around humans but they live immersed in the human socio-linguistic environment,” she said. “They are exposed to speech on an everyday basis. They have toys. They live in the human physical environment as well. They are part of our family.”

Dogs are good at responding to instruction words such as “sit,” or “come” and while some dogs have learned hundreds of object names, others struggle to do so, Boros said. The research team wanted to find out why that is.

The team wanted to look into how much dogs understand the human language.

“We want to know whether animals can have certain language skills which are present in humans,” said co-lead author Magyari. 

How researchers, dogs and the humans who love them made this happen

The researchers looked into canine cognition using methods previously used on infants. 

Owners brought their dogs to the lab, along with toys they normally play with. Researchers recorded the dog owners referring to different toys.

The dogs were separated from their owners and sat across from them and peered at the owners through a window, almost like a television with a green screen.

Researchers first played the pre-recorded messages from the dog owners, then the dog owners held up toys for the dogs to see in live action. Sometimes they tricked the dog by playing one message but showing the dog a different toy.

The researchers said they can infer from the dogs’ brain activity that the canines were expecting to see the object named by the owner. 

When that didn’t happen, “it violated that expectation” resulting in different brain activity, Magyari said. 

Limitations of the study

Magyari said the study has its limitations. First, this is just one study and researchers need to conduct more.  

The researchers also don’t know how this skill developed in dogs. It could be domestication or evolution. 

The researchers are happy the experiment was a success because now, researchers can do different versions of this study to find out more, she said.

“I think there are still many other open questions,” Magyari said.

Saleen Martin is a reporter on USA TODAY's NOW team. She is from Norfolk, Virginia – the 757. Follow her on Twitter at @SaleenMartin or email her at [email protected] .

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Animals (Basel)

What Kinds of Dogs Are Used in Clinical and Experimental Research?

Associated data.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Simple Summary

The objective of this study was to evaluate the signalment of dogs used in veterinary research in six different specialties. In total, 150 randomly chosen clinical studies (25 studies per specialty) published between 2007 and 2019 were evaluated for the breed, sex, neuter status, age, and weight information of the dogs used. Breed information was given for 5.7% of the included animals. Beagles were used 1.9% of the time, which was a less significant role in research than we expected. Information about the sex of the dogs was lacking for 16.2% of the included animals, while age and weight information were missing for 22.7 and 32.7%, respectively. The neuter status was not given in 38.7% of the clinical studies. The results show deficits in the reporting of demographic data for the dogs. The need for an improvement in the documentation and/or reporting of animal signalment is obvious and should be addressed by authors, reviewers, and journal editors in the future.

Background: Dogs are widely used in research to answer questions about canine or human conditions. For the latter, research dogs are often used as models, since they are physiologically more similar to humans than other species used in research and they share similar environmental conditions. From a veterinary perspective, research findings are widely based on academic research, and thus are generated under experimental conditions. In that regard, the question arises: do the dogs used for research adequately represent the dog population seen in veterinary practice? It may, for example, be assumed that Beagle dogs are often used as experimental animals. The objective of this study was to evaluate the signalment of dogs used in veterinary research. Furthermore, we aimed to assess other relevant criteria regarding the validity of clinical trials in the context of six different veterinary medicine specialties: cardiology, internal medicine, neurology, orthopaedics, reproduction, and surgery. Methods: A literature search was conducted and 25 studies per specialty were randomly selected. The breed, sex, neuter status, median age, and median weight of the dogs used for clinical studies ( n = 150) published between 2007 and 2019 were evaluated. Results: In total, 596,542 dogs were used in the 150 trials. Breed information was given for 33,835 of these dogs (5.7%). Of the latter, 1.9% were Beagles. Nine clinical trials exclusively used Beagles. The most frequently used breeds were German Shepherds (7.3%), Labrador Retrievers (6.7%), and Golden Retrievers (4.7%). The major reporting deficits found were missing breed specification in 25.3% of the articles; missing information about the sex of the dogs in 16.2%; missing age and weight information in 22.7 and 32.7%, respectively; and missing neuter status in 38.7% of the clinical studies. The median sample size was 56 (Q1:29; Q3:365) dogs. Conclusions: The presented project revealed that Beagle dogs represent only a small proportion of dogs in veterinary research. Based on the evaluated publications, it seems that some relevant dog attributes differ between the specialties. The results, however, show deficits in the reporting of demographic data for the dogs. The need for an improvement in the documentation and/or reporting of animal signalment is obvious and should be addressed by authors, reviewers, and journal editors in the future.

1. Introduction

In medical research, dogs have been and still are widely used as testing animals [ 1 ], although worldwide, actual and precise figures are not easy to retrieve because relatively few countries collate and publish research-animal statistics [ 2 ]. It has been estimated that 79.9 million animals, out of which 207,724 were dogs, were used worldwide for experimental or scientific purposes in 2015 [ 3 ]. This was a 36.9% increase in the equivalent estimated figure of 58.3 million animals in 2005 [ 2 , 3 ]. The most widespread use of experimental animals occurs in China, followed by Japan and the United States [ 4 ]. Around 800,000 laboratory animals were used in 2019 in the US, of which 7% (56,000) were dogs [ 5 ]. In the EU from 2017 to 2018, the number of animals used in research decreased by five per cent to around nine million animals, of which 0.3% (27,000) were dogs and cats [ 6 ]. The purposes of research projects on dogs include gaining basic biological knowledge; answering questions regarding human health by using dogs as models for the development of drugs, diagnostic tests, vaccines, and medical devices; and answering questions directly linked to canine physiology or diseases [ 7 ]. Dogs are often preferred as models for human conditions because they are physiologically and clinically more similar than other species such as mice [ 8 ], and pet dogs also share the environmental conditions of their owners. In addition, the domestic dog, canis familiaris , reportedly bears over 450 diseases; approximately 360 of these are analogous to human diseases [ 9 ]. These analogous conditions include diabetes, cancer, epilepsy, eye diseases, and autoimmune diseases, not to mention the high numbers of rare monogenetic diseases [ 8 ]. It has been stated that the most suitable and frequently utilized breed in clinical trials is the Beagle. These dogs are commonly kept in kennels in homogenous groups based on age, weight, sex, and neuter status in the research facilities of universities or pharmaceutical companies [ 10 ]. Beagles are medium-sized and have a short coat and an even temperament, which makes them particularly suitable for medical research in contrast to other breeds [ 11 ]. These advantages make it easier to standardise specific and relevant conditions in research settings and help keep the costs of the experiment lower [ 12 ].

Worldwide, more than 354 dog breeds are registered at the FCI (Federation Cynologique Internationale) [ 13 ]. Searching for “Beagle dog” in the PubMed database for results between 2007 and 2019 ( http://www.pubmed.gov , accessed on 21 May 2022) leads to more than 4790 results. When performing the same search with the breed “Labrador dog”, which has been the most popular breed in the US since 1991 [ 14 ], the results are much less (1008 results). These figures support the assumption that Beagles may be the most utilized research dogs by far.

Both dog breeds can be found in the top ten list of the most popular dog breeds in the US [ 15 ] ( Table 1 ). Beagles often serve as models for research focusing on human health, such as the toxicology of medications [ 16 ]. Research on Beagles has led to various relevant findings for human medicine, such as information about infections with Helicobacter Pylori [ 7 ], or better insights into the effects of the frequency of tooth brushing [ 17 ].

Top 10 dog breeds in the US in 2019 and the number of publications found in PubMed for each breed (search date: 21 May 2022).

Depending on the research question, alternatives for the use of research dogs may be the enrolment of client-owned dogs or the retrospective evaluation of medical records. Findings from these approaches might even better represent the heterogeneity of dogs seen in daily practice and better address real clinical conditions. However, depending on the research question, a high level of heterogeneity regarding breed, age, sex, weight, neuter status, housing, feeding, and other parameters in a study population may lead to a significant influence of confounders [ 18 , 19 ]. If potential confounders are not taken into account during the analysis and interpretation of research results, this may lead to biased outcomes and conclusions. There is evidence that the sex, weight, age, breed, and neuter status of the dogs that take part in a clinical trial are important when it comes to different conditions and diseases, such as joint disorders [ 20 , 21 ], metabolic conditions [ 22 ], and periodontal diseases [ 23 ]. In addition, breed differences significantly affect the incidences of specific diseases such as pyometra [ 24 ], dilatative cardiomyopathy [ 25 ], and granulomatous colitis [ 26 ], as well as sex-steroid-influenced diseases such as diabetes [ 27 ] and hyperadrenocorticism [ 28 ].

The objective of this study was, therefore, to evaluate what kinds of dogs were utilized in clinical trials. We aimed to assess and compare information about the dogs, including the breed, sex, age, weight, and neuter status in the context of six different veterinary medicine specialties. In addition, we aimed to evaluate the extent to which Beagles have been used in veterinary research.

2. Material and Methods

A literature search in the databases PubMed ( http://www.pubmed.gov accessed on 31 October 2020) and CAB Abstracts ( https://www.cabdirect.org/ , accessed on 31 October 2020) was conducted on 31 October 2020. The same literature search and selection process for articles was used and described in detail in another research project [ 29 ] assessing different literature parameters. In brief, the following search keywords were used: clinical trial AND dogs AND specialty. For each search procedure, “specialty” was replaced by cardiology, internal medicine, neurology, orthopaedics, reproduction, or surgery.

Publications had to be in the English or German language and published in or between the years 2007 and 2019. Case reports or case series with a number of animals lower than n < 10, opinions, clinical experiences, and abstracts with less than 500 words were excluded. Studies or case reports without statistical analysis and studies on other species, such as humans or cats, were also excluded. In addition, in vitro studies were not included. From the final 150 articles, 134 articles were accessed via online databases, nine papers were retrieved in the veterinary library at the University of Berlin, and two articles were obtained via inter-lending from other libraries. As a first step, the literature was evaluated using a slightly modified version of the checklist developed in 2010 by Arlt and Heuwieser [ 30 ]. The results have been published in a previous article [ 29 ]. In addition to the validated checklist, the following items were assessed for the presented project: number of dogs, number and type of dog breeds, gender, neuter status, median weight, and median age.

Statistical Analysis

All statistical analyses were conducted in IBM SPSS for Windows (Version 24.0; SPSS Inc., Munich, Germany). Categorical data were presented descriptively as raw numbers and percentages. To identify differences between the specialties, the non-parametric Mann–Whitney U test was used as indicated by the distribution. The statistical significance was set at p < 0.05.

From the 150 examined studies, 108 (72.0%) were prospective and 42 (28.0%) were retrospective. Considering the study design, 91 publications were classified as interventional studies (60.7%) and 59 were observational studies (39.3%). In total, 596,542 dogs were used in the 150 clinical studies assessed in this project. For one study, the number of dogs was not specified; instead, the number of limb fractures over a given period of time was reported. For statistical reasons, we set the number of limb fractures equal to the number of dogs. In 100 articles (66.0%), the breeds of all enrolled dogs were specified, leading to 33,835 dogs with breed information. Out of the remaining 50 trials, breed information was not given at all for 38 studies and was incomplete for 12 studies. Most studies with no or incomplete breed information were retrospective. In several studies, the breed was not specified for all dogs included, but only for the numerical top ten breeds. Analysing breed information in the 12 studies with incomplete data led to another 7792 dogs with given breeds and 5384 with missing information. Overall, breed information was available for 41,627 dogs, which was 6.9% of the overall number of dogs. The overall median number of dogs used in each of the studies was 56 (Q1:29; Q3:365). Retrospective studies had a larger number of included animals (Median: 62; Q1:35; Q3:384) than prospective ones 24 (Q1:13; Q3:41), p < 0.05. Out of the dogs with known breeds, 643 dogs (1.5%) were Beagles. In nine studies, the dog population consisted of Beagles only. The median sample size of these studies was 12 dogs (Q1:12; Q3:24). These studies included five experimental studies (two related to human research) and four clinical trials (one related to human research) carried out to determine the effectiveness or administration route of drugs. All studies took place under laboratory conditions. One article was published in each of the years 2007, 2008, 2014, and 2015, two articles were published in 2017, and three were published in 2018. Furthermore, in 32 studies, Beagle dogs were used among dogs belonging to other breeds. In Section 3.4 , more information about the dogs used in experimental trials is given. Out of 150 studies, 16 (10.6%) were related to human research using dogs as a model, and three of these used study populations consisting of Beagles only.

Within the individual specialties, the number of Beagles used as clinical trial dogs was heterogenous ( Table 2 ). The greatest numbers of Beagles were used in internal medicine and surgery studies (each n = 8, 32.0%).

Use of Beagles as trial dogs within six veterinary specialties in 150 clinical trials (25 per specialty).

For dogs with a known breed, the proportion of Beagles differed between the specialties. In internal medicine, we found the highest proportion of Beagles, with 12.9%. Orthopaedics followed with 5.6% and cardiology with 4.4%. The proportion was smallest in reproduction, with 0.4% ( Table 3 ).

Numbers and proportions of Beagle dogs utilized in clinical trials within six veterinary specialties ( n = 150).

The highest amount of missing breed information was found in internal medicine (48.0%), followed by orthopaedic studies (40.0%), reproduction (36.0%), cardiology (28.0%), and neurology (20.0%). Only a portion of the dogs used in the 31 prospective studies (20.6%) and the 19 retrospective studies (12.6%) had their breeds specified. The median number of breeds was four for prospective studies (Q1:1; Q3:10), while retrospective studies had a median number of sixteen breeds (Q1:3; Q3:27). For the different specialties, the median number of breeds was one for cardiology (Q1:1; Q3:9.5), ten for internal medicine (Q1:4; Q3:15), nine for neurology (Q1:1; Q3:21), five for orthopaedics (Q1:1; Q3:16), four for reproduction (Q1:1; Q3:8.5), and nine for surgery (Q1:4; Q3:13). For 40 (26.7%) studies, there was only one breed utilized; 14 (56.0%) of these studies were cardiology trials with mostly mongrel dogs. Table 4 lists the US top ten breeds of 2019 and the number of dogs belonging to each breed that have been used in clinical veterinary research worldwide. The popularity of the dog breeds in Europe are similar [ 31 ]. These dog breeds seem to play an important role in research, since they make up around 25.2% of all study dogs with given breed information. Interestingly, despite their popularity, Bulldogs were not extensively used in the studies selected in this project.

Numbers and proportions of dogs belonging to the US top ten breeds of 2019 used in research ( n = 150).

3.1. Sex of the Dog Population

In total, the sex of the dogs was specified in 83.8 % ( n = 150) of the trials ( Figure 1 ). Dogs of both sexes were used in 89 studies (59.3%), solely females in 25 trials (16.7%), and only male dogs in 11 studies (7.3%). Both sexes were predominantly mixed in 22 neurology studies (88.0%) and 20 internal medicine studies (80.0%). In studies relating to reproduction, dogs of only one sex were used in most trials ( n = 20, 80.0%). In seventeen studies, female dogs were used (68.0%), and in five studies, both female and male dogs were used. In contrast to the other specialties, the sex of the dogs was determined for all 25 reproduction studies. In around 40% ( n = 10) of the studies on cardiology, the sex of the dogs was not documented, followed by six orthopaedic studies and six surgery studies with an unknown sex for the dogs used (24.0%).

Sex of the dogs utilized in 150 clinical trials within six veterinary specialties (25 each).

Spay and neuter status of the study population. The spay and neuter status of most dogs used in the 150 studies was not specified in 38.7% of the studies ( Figure 2 ). In four studies (2.7%), all dogs used in the trial were neutered. In 34.7% of the studies, both neutered and intact dogs were used, and in 24.0% of the studies, all dogs were intact. Except for studies in the field of reproduction, the neuter status was not specified as an inclusion or exclusion criterion. In 23 studies (92.0%) belonging to the reproductive field, intact dogs were enrolled ( p < 0.05). In 18 neurology studies (72.0%) and 13 internal medicine (52.0%) studies, both neutered and non-neutered dogs were used. The spay and neuter status of the dogs used was not specified in 16 (64.0%) studies each on cardiology and orthopaedics.

Spay and neuter status of the dogs utilized in 150 clinical trials within six veterinary specialties (25 each).

3.2. Age of the Study Population

Overall, age information was not given in 34 (22.7%) studies ( Figure 3 ). The median age of all dogs with given information was 5.0 years (Q1:2.4; Q3:6.9), the minimum age was 2 weeks, and the maximum age was 12.5 years. In six studies (4.0%), dogs with a median age of under one year were used. Age information was missing most often in cardiology studies ( n = 11, 44.0%), followed by reproduction studies ( n = 8, 32.0%). The median age for dogs enrolled into cardiology studies was 9.0 years (Q1:6.4; Q3:11.7), internal medicine was 6.0 years (Q1:4.9; Q3:7.6), neurology was 4.4 years (Q1:2.7; Q3:6.9), orthopaedics was 4.8 years (Q1:2,4; Q3:5.6) reproduction was 3.3 years (Q1:2; Q3:5.2), and surgery was 4.0 years (Q1:2.1; Q3:5.45).

Age (in years) of the dogs utilized in 150 clinical trials within six veterinary specialties (25 each).

3.3. Weight of the Dogs

Weight information was not given in 49 publications (32.7%) ( Figure 4 ). The overall median weight was 20.3 kg (Q1:10.8; Q3:27.2), the lowest median weight of a study population was 3.7 kg, and the highest median weight of the dogs used within one study was 50.3 kg. Weight information was missing most often in reproduction studies ( n = 17, 68.0%), followed by neurology studies ( n = 12, 48.0%), cardiology and orthopaedics studies (each n = 6, 24.0%), and surgery and internal medicine (each n = 4, 16%). The median weight of the dogs used in cardiology studies was 14.5 kg (Q1:9.3; Q3:22.0), internal medicine was 20.2 kg (Q1:9.3; Q3:23.7), neurology was 19.3 kg (Q1:12.5; Q3:20.6), orthopaedics was 31.2 kg (Q1:25.6; Q3:33.8), reproduction was 17.8 kg (Q1:11.8; Q3:21.1), and surgery was 22.0 kg (Q1:10.7; Q3:25.2). One study in the field of surgery used dogs with an average weight of over 50 kg. The weight differences of dogs used in orthopaedics compared with neurology and cardiology were significant ( p < 0.05).

Weight (in kg) of the dogs utilized in 150 clinical trials within six veterinary specialties (25 each).

3.4. Origin of Study Population and Overall Number of Animals

Information on the ownership of the dogs and the number of dogs used for the different studies has already been presented in a previous paper [ 29 ]. The dogs were either experimental animals owned by the research institutions (18.7%), privately owned (76.0%), mixed (0.7%), or the origin of the dogs was not specified (4.7%). For research in the field of cardiology, experimental animals were used in 11 of the articles (44.0 %). The dogs used for the 28 experimental studies had a median age of 2.3 (Q1:2; Q3:4.6) years and a median weight of 15.5 kg (Q1:11.9, Q3:22.5). The neuter status of the dogs was not given in twenty articles, seven study populations were not neutered, and one had a population with a mixed neuter status. For the experimental studies, the gender of the dogs was unknown for eleven study populations, only female dogs were used in five studies, only males in seven studies, and a mixed population in five studies. The median number of dogs used for experimental studies was 21.0 (Q1:11.5, Q3:24.3) and the median number of breeds was 1.0 (Q1:1.0; Q3:1.0). In 27 out of the 28 experimental studies, the breed of the dogs was specified. The study population of nine studies consisted of cross-bred dogs and seventeen study populations were purebred dogs. Most of the experimental studies (n = 9) consisted of Beagles only, followed by three studies with a mixed study population of Beagles, Labradors, Rottweilers, and German Shepherds, and one each of Foxhounds, Sheepdogs, Pitbull Terriers, and Coonhounds.

The number of animals was given in most of the articles in all different specialties. For two studies—one on reproduction and one on orthopaedics—the authors did not specify the number of enrolled animals. The median sample size for all the included studies was 31 dogs (Q1:16; Q3:64). For the different specialties, the median sample size was 32 for cardiology (Q1:22; Q3:207), 31 for internal medicine (Q1:14; Q3:53), 33 for neurology (Q1:15; Q3:56), 36 for orthopaedics (Q1:19; Q3:95), 35 for reproduction (Q1:18; Q3:74), and 25 for surgery (Q1:16; Q3:40).

4. Discussion

This study focused on the assessment of the signalment, such as breed, sex, neuter status, age, and weight, of dogs used in controlled clinical trials. These data were analysed with consideration of the veterinary specialty that each publication belonged to. The results of a critical appraisal of the quality of the clinical studies have been published in an earlier paper [ 29 ]. The articles were randomly selected from a list generated after a literature search in two relevant databases. There were sixteen studies using dogs that served as model for human health, but only three studies used Beagles as a study population. Interestingly, nine of these studies belonged to the specialty of cardiology.

The proportion of Beagles in our literature sample was less than we expected and less than several authors have claimed [ 10 ]. A reason could be that some studies focus on conditions that occur naturally in specific breeds, such as diseases related to brachycephaly. In 2020 in the UK, 99% of dogs used in experimental research (medical and veterinary) were Beagles [ 32 ]. In other European countries, breed information is not given in the statistics about animal testing [ 6 ]. Obviously, the Beagle has not been a predominant breed used for clinical veterinary studies in the past years. In fact, the breed of dogs used in research more or less reflects the most popular breeds, with the exception of Bulldogs. The reason for the latter remains open and should be further investigated. Nevertheless, the presented data may imply a slight increase in the use of Beagles over the last years, but the number of the assessed articles is much too small to conclude a trend. On the contrary, the number of dogs used as experimental animals seems to be declining [ 5 ], albeit a lack of comprehensive data on the worldwide use of dogs in experimental research and clinical trials. The use of Beagles and dogs in general for basic research and research related to human medicine may have become less popular in the past decades because of an increasing awareness of animal welfare [ 4 ]. Just recently, a Beagle-breeding company was inspected and found itself confronted with allegations of severe animal welfare violations, which were reported in different media [ 33 ]. Over the years, the European Union, Canada, the United States, and several other countries have introduced laws to regulate the use of laboratory animals for medical research after consulting the main stakeholders [ 4 ]. In addition, laboratory dogs are often rehomed nowadays into private households after their use in research [ 34 , 35 ]. However, it has been claimed that they still experience an extreme change in their life situation because they leave their familiar, limited environment in the research facility and encounter a multitude of animate and inanimate stimuli in their new home [ 36 , 37 ]. For 50 studies (33.3%), the amount of missing or incomplete information about the breeds used for the clinical studies is relatively high, which shows that this kind of information seems to be considered unimportant by authors, reviewers, and editors. However, since breed may be a relevant confounder, this information is essential data that needs to be given according to the STROBE statement [ 38 ]. The median number of breeds used for the clinical trials does not differ significantly between the veterinary specialties. It is noticeable that most of the cardiology studies were conducted with experimental dogs and most of the studies consisted of just one breed. The evaluation of the other parameters showed that, for most trials, both female and male dogs were used, meaning that sex was not set as an inclusion or exclusion criterion. The sex of all dogs was only given in reproduction trials, and more than 60% of these trial dogs were female. It is plausible that studies on gynaecology usually relate to conditions found in intact bitches. For the other veterinary specialties, the sex of the dogs seems to be considered unimportant, and this might indeed be the case. Similar findings have been documented regarding the neuter status. More than 90% of the dogs used for reproduction trials were intact, while in the studies belonging to the other veterinary specialties, both intact and neutered dogs were often used. It can be concluded that the neuter status seems to be considered an irrelevant factor or confounder for most research questions beyond reproduction. It is noticeable that the median age of the dogs used for cardiology was nearly double that of the ages of the dogs enrolled into studies belonging to the other veterinary specialties. This might reflect the fact that cardiac diseases tend to occur or be diagnosed in older dogs. The median age of the dogs used for the reproduction studies was the lowest. This may be related to the optimal breeding age of bitches. The median body weight of the dogs was highest in the orthopaedic studies, at around 30 kg. It is common knowledge that especially large and heavy dogs have a higher risk of suffering from joint diseases [ 20 ]. For other disorders, factors such as age and weight may also play a role concerning a studied condition. Therefore, this information should be given in scientific articles. The presented review has revealed documentation and reporting deficiencies. Based on the presented data, it is not possible to judge whether the study methodology or reporting is better or worse in specific veterinary specialties. In fact, it seems that even if some aspects are better presented in one specialty, other factors are missing more often in the same specialty. Similar shortcomings have been described by Reynolds et al. [ 39 ] and others [ 29 , 40 , 41 ]. Demographic data for dog populations used in clinical trials is very important, since it is needed to draw sound conclusions and extrapolate the findings [ 42 ]. In addition, comprehensive information about the used animals is important for readers to assess whether the given scientific information should be applied in an actual case. It has been proven that the signalment such as age, sex, and breed are highly relevant aspects for examining the prevalence of conditions and the interpretation of various study outcomes. Besides the examples mentioned earlier, age is a key factor for the outcomes of electrocardiographic exams [ 43 ], fertility [ 44 ], behaviour [ 45 ], and several canine diseases [ 46 , 47 , 48 ]. Belic et al. found that the sex of a dog plays an important role in biochemical markers for the bone turnover [ 49 ]. Anatomical or hormonal differences between male and female or neutered and intact dogs can also have an impact on the prevalence of diseases [ 50 , 51 , 52 ] or the outcome of clinical studies [ 53 ]. In addition, it has been shown that the breed has an effect on renal size [ 54 ], and genetic and phenotypic differences across dog breeds have an influence on the safety and efficacy of pharmaceutical substances and their doses [ 55 ] and the occurrence of genetic diseases [ 56 ] such as MDR1 mutations [ 57 ]. Several studies on the methodological and reporting quality of clinical trials in veterinary medicine have been published in the past years. Limitations have been criticized, such as small numbers of included animals; a lack of sufficient reporting on the specifications of animals, diagnoses, and treatments; and undocumented inclusion and exclusion criteria [ 29 , 42 , 58 ]. Our results show that relevant reporting deficiencies in clinical studies were found and essential information about the dogs used in clinical trials was missing. This may considerably limit the validity of research results. More attention should be paid to reporting guidelines, such as the STROBE statement for observational trials or the CONSORT statement for randomised studies [ 59 ], as they have been developed to improve the quality of scientific articles.

The estimated number of dogs used worldwide for medical research is still high. Attempts have been made to replace laboratory animals by in vitro and in silico methods [ 1 , 60 , 61 , 62 ]. For laboratory animals in most countries, the consideration and implementation of replacement, refinement, and reduction (3Rs) strategies, proposed by Russell and Burch in 1959, is mandatory [ 63 ]. For some purposes, animal testing has been forbidden by some authorities. For instance, the use of animals for cosmetic testing has been prohibited in Europe as of March 2013 [ 64 ].

Limitations

There are some limitations to the present study, such as the relatively small sample size of 25 articles per specialty. Since this number per specialty led to a total number of 150 articles, which was eligible for a throughout assessment, the inclusion of more literature was not possible within this project. In that regard, it may be worthwhile to re-evaluate some of the presented specialty-specific findings on a larger scale with a greater number of articles.

Furthermore, the investigation and evaluation of the studies was done by just one non-blinded researcher, which may have led to a biased interpretation. This approach, however, has been used in several other studies before [ 29 , 30 , 41 , 65 ].

5. Conclusions

The results of our literature review concerning the kinds of dogs that are used in veterinary research indicate that we are widely not able to give a sufficient answer. The presented project revealed that Beagle dogs represent only a small proportion of dogs in clinical veterinary research. It seems that Beagles are used much more in experimental research, but this should be investigated in future analyses including more experimental research reports. The results of this study are furthermore in accordance with previous findings, and reflect once more that essential information about the dogs used in clinical trials is missing. Some parameters, such as body weight and neuter status, vary significantly between specialty-specific studies. Authors, reviewers, and journal editors should pay more attention to the reporting of basic information about the animals enrolled in veterinary research and should follow the guidelines for the specific study type.

Acknowledgments

The authors thank Peggy Haimerl for participating in the pre-test and for her support and advice. The publication of this article was funded by Freie Universität Berlin.

Funding Statement

The authors received no external funding for this work.

Author Contributions

E.S. and S.P.A. drafted and revised the manuscript. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable. Since no animals were used in the context of this manuscript, no ethical permit was required.

Informed Consent Statement

Not applicable.

Data Availability Statement

Conflicts of interest.

None of the authors of this paper have a financial or personal relationship with other people or organisations that could inappropriately influence or bias the content of the paper.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.