Serotonin and Dopamine Blood Levels in ADHD-Like Dogs

Simple Summary Attention deficit hyperactivity disorder (ADHD) is a relatively common neurodevelopmental disorder mainly affecting children and adolescents. ADHD is associated with significant social stigma, psychosocial adversities, and reduced working abilities. As with humans, dogs can suffer from ADHD-like behaviors, but to date, there are few studies on this condition in the canine species. Therefore, the present study analyzed the levels of serum serotonin and dopamine in dogs with signs that are similar to the ones presented in human patients with ADHD (compared to control dogs) that were assessed clinically and through different behavior scales. The results obtained indicate that both neurotransmitters have a tendency to be observed in lower concentrations in dogs presenting behavioral signs consistent with ADHD-like disorders. This finding may be useful for contributing to the study of the diagnosis and treatment of this disorder in dogs. ADHD-like associated behaviors can disrupt the normal coexistence of the dogs in their homes and contribute to abandonment, which continues to be a major social and animal welfare problem today. Abstract As with humans, dogs can suffer from attention deficit hyperactivity disorder-like (ADHD-like) behaviors naturally and exhibit high levels of hyperactivity/impulsivity and attention deficit problems, making the domestic dog a potential animal model for ADHD. ADHD has a very complex pathophysiology in which many neurotransmitters are involved, such as serotonin and dopamine. The aim of the study was to evaluate serum serotonin and dopamine levels in dogs with ADHD-like symptomatology. Fifty-eight dogs were studied, of which, thirty-six were classified as ADHD-like after physical and behavioral assessments. Additionally, the dogs’ owners performed a series of scientifically validated questionnaires which included C-BARQ, the Dog Impulsivity Assessment Scale, and the Dog-ADHD rating scale. Serum from every animal was collected after the behavioral assessments and analyzed with commercial ELISA tests for serotonin and dopamine determination. Kruskal–Wallis tests and Lasso regressions were applied to assess the relationships between both neurotransmitters and the ADHD-like behaviors (as assessed by clinical evaluation and through the different questionnaires). The dogs clinically classified as ADHD-like showed lower serotonin and dopamine concentrations. Further, serotonin and dopamine levels were also linked to aggression, hyperactivity, and impulsivity. Decreased serotonin concentrations were also related to fear, attachment, and touch sensitivity. Finally, it must be noted that our data suggested a strong relationship between serotonin and dopamine and ADHD-like behaviors.


Introduction
ADHD (attention deficit hyperactivity disorder) is one of the most frequently diagnosed syndromes in child psychiatry. The key symptoms (inattentiveness, impulsivity, and hyperactivity) deteriorate the relationships of these children both in their families and with serotonin levels among hyperactive and ADHD children [24,29,30]. In fact, a chronic deficiency in available serotonin may contribute to the clinical symptoms of ADHD [31]. Interestingly, lower levels of urine serotonin and dopamine were found in dogs with higher impulsivity [32]. In addition, aggressive and impulsive dogs showed lower levels of blood serotonin [33][34][35][36]. Nevertheless, there are no previous studies assessing serotonin or dopamine levels in ADHD-like dogs.
Therefore, the aim of the present study was to evaluate serum serotonin and dopamine levels in ADHD-like dogs (vs control dogs) as assessed by clinical evaluation and in dogs assessed through different validated questionnaires.

Surveyed Animals and Data Collection
The study was carried out at the Rof Codina Veterinary Teaching Hospital (VTHRC), which is part of the Veterinary Faculty of Lugo (the referral center for veterinary clinics in northwest Spain), with the collaboration of two other leading behavioral medicine services, Lar do Belelle and Etoloxía, also in northwest Spain. A total of 58 dogs were included in the study. They were recruited through cases that came to behavioral medicine services because of hyperactivity, impulsivity, or attention deficit problems and through an active search of dogs with ADHD-like signs, and the control dogs were recruited through social media networks such as Facebook, Instagram, and LinkedIn. We also shared an e-mail chain, starting with our personal contacts. Therefore, 40 of the 58 dogs had attended consultation at the VTH RC, with 10 at the Lar do Belelle center and 8 at the Etoloxía center. During consultation, we collected data on gender, entire or neutered, age (months), age at acquisition (months), weight (kg), activity pattern (whether the dog performed some activity such as agility, dog jogging, dog trekking, dog biking, mushing, or other), whether the owner had had other dogs previously, and habitat (rural or urban). Evaluations of the dogs were carried out by a veterinarian specializing in behavioral medicine.

Dog Assessment
As mentioned, all the dogs were examined by a veterinarian specializing in behavioral medicine. A complete history and behavioral and physical examinations (with the collection of blood samples) was carried out. ADHD-like were suspected in cases that presented hyperactivity, impulsivity, or attention deficit symptoms (a list of different symptoms of the ADHD-like dogs are described in Table 1) [8,25,37].  [8,9,25,37]. Based on the above and on the Canine Behavioral Assessment and Research Questionnaire (C-BARQ) [38][39][40], 5 dogs with suspected attention deficit related to phobias, hyperactivity that could be associated with a lack of exercise, and impulsivity that could originate from previous learnings, as well as patients with organic pathologies (including chronic pain), were not included in the final sample of 58 animals.
In addition, all the dogs' owners completed a series of scientifically validated questionnaires in addition to the C-BARQ: Dog Impulsivity Assessment Scale (DIAS) [41,42] and the Dog-ADHD rating scale, owner's version [8].
The C-BARQ questionnaire consists of 100 questions describing the different ways in which dogs typically respond to common events, situations, and stimuli in their environments. Responses are grouped into 14 behavioral traits: stranger-directed aggression, owner-directed aggression, dog-directed aggression, dog-directed fear, familiar dog aggression, trainability, chasing, stranger-directed fear, nonsocial fear, separation-related problems, touch sensitivity, attachment/attention seeking, excitability, and energy. All traits are expressed on a 0 to 4 scale in which 0 indicates no sign of such behavior and 4 indicates a severe form of the behavior.
The DIAS is an owner-based questionnaire with 18 statements concerning their perceptions of the behavioral tendencies of their dog, with individual item responses scored on a 1-5 scale. The scores were calculated for the overall questionnaire (OQS) and for each of the three subscales that compose it: "Behavioural Regulation", "Aggression & Response to Novelty", and "Responsiveness" [32,[41][42][43].
The Dog-ADHD rating scale, owner version (Dog ARS) was developed on the basis of a validated human ADHD questionnaire. In the Dog ARS, six items are designed to refer to attention deficits in a dog (Subscale I) and seven are designed to measure the level of motor activity and impulsivity (Subscale II). The owners have to choose answers from the presented alternatives representing different frequencies of how often the statement is true for their dog on a four-level scale (never, rarely, often, or very often) [8].
The dogs were treated according to the European and Spanish legislation on animal protection (Directive 86/609/EEC, Real Decreto 1201/2005), and the Ethical Committee of the VTH Rof Codina approved the experiments and procedures.

Serotonin and Dopamine Measurement
Serum from every animal was collected after the behavioral assessment. Blood samples were obtained by venipuncture using tubes without anticoagulant from the jugular, cephalic, or saphenous vein. After that, they were centrifuged and frozen at −80 • C until analysis.
The serum samples were analyzed with two commercial ELISA tests (the Serotonin High Sensitive ELISA, DLD Diagnostika GMBH, Hamburg, Germany), and the Canine Dopamine DA ELISA Kit, MyBioSource, San Diego, CA, USA). These are competitive ELISA kits for the quantitative determination of serotonin and dopamine, respectively.
The analyses were performed following the recommendations and procedures indicated by the manufacturer for serum samples. According to the test, concentrations were expressed in ng/mL.

Statistical Analyses
Statistical analyses were performed using R Statistical Software (Supplementary Table S1). Initially, the scores obtained for the studied animals in the different traits for the DIAS, Dog ARS, and C-BARQ questionnaires were transformed into binary variables, taking as cut-off points for each trait the median of the study population for that trait. This allowed for dividing the population into two groups of similar sizes. Subsequently, the differences in the serotonin and dopamine concentrations for the suspected ADHD dogs, based on the behavioral examinations (yes/no) and the different DIAS, Dog ARS, and C-BARQ tests (scores under or over the median), were examined by Kruskal-Wallis tests.
Finally, a Lasso regression (glmnet package, https://cran.r-project.org/web/packages/ glmnet/index.html, accessed on 28 June 2021) was applied to evaluate the effects of sero-tonin and dopamine levels on the suspected diagnosed ADHD dogs (yes/no) and the different DIAS, Dog ARS, and C-BARQ tests (as mentioned, and transformed into the binary variables of scores above or over the median). Gender, reproductive status (entire or neutered), age (months), age at acquisition (≤2 months or >2 months of age), weight (kg), activity pattern (whether the dog performed some activity such as agility, dog jogging, dog trekking, dog biking, mushing, or other), whether the owner had had other dogs previously, and habitat (rural or urban) were also included in the models as independent control variables. One model was performed for the ADHD-like suspected symptoms and one for each DIAS, Dog ARS, and C-BARQ test.
A K-fold cross-validation was performed to determine the lambda value that produced the lowest test mean squared error. The main advantage for the Lasso regression was that the model performed automatic feature selection to decide which variables should and should not have been included on their own. It helps to reduce overfitting and it is useful for variable selection when there are several independent variables that may not contribute to the outcome.

Results
After the behavioral examinations, 36 of the 58 dogs (62.0%) were classified as ADHDlike (presenting hyperactivity, impulsivity, or attention deficit symptoms). The remaining ones were used as control dogs (no ADHD-like symptoms). Descriptive features of the studied population (total and stratified by ADHD-like diagnosis) are shown in Table 2. The mean serotonin and dopamine serum concentrations in the dogs with suspected diagnoses as being ADHD-like (vs. the not suspected ADHD-like dogs) and the dogs with scores over and under the median for the different DIAS, Dog ARS, and C-BARQ traits evaluated are provided in Table 3. In the univariate approach (Kruskall-Wallis test), the differences between the groups for the serotonin serum concentrations were significant for the suspected ADHD-like dogs, the DIAS total score, the DIAS behavioral regulation score, the DIAS aggression score, the Dog ARS activity-impulsivity score, and the C-BARC owner-directed aggression, dog-directed fear, and non-social fear scores. In the case of dopamine, significant differences were observed for the dogs with a suspected diagnosis for being ADHD-like, for the ARS activity-impulsivity score, and for the C-BARC energy score. The p-values obtained from the Kruskall-Wallis test are provided in Appendix A. Box diagrams of the serotonin and dopamine levels for the dogs with suspected diagnoses of being ADHD-like (vs the control dogs) and the dogs with scores over and under the median for the different DIAS, Dog ARS, and C-BARQ traits are shown in Figure 1. Table 3. Mean serotonin and dopamine serum concentrations (ng/mL) of the dogs with suspected diagnoses for being ADHD-like (vs. the not ADHD-like dogs) and the dogs with scores over and under the median for the different DIAS, DOG ARS, and C-BARQ traits (different superscripts indicate significant differences). The mean scores obtained for the dogs with scores over and under the median for the different traits are provided, along with the standard deviations (SDs).  The multivariate Lasso regression indicated that the dogs with lower serotonin and dopamine serum concentrations were more likely to be diagnosed as being ADHD-like. The results also suggested that gender, neuter status, and age were significant contributors to the prevalence of ADHD-like symptoms (with higher odds in male, entire, and younger dogs) ( Table 4). case of dopamine, significant differences were observed for the dogs with a suspected diagnosis for being ADHD-like, for the ARS activity-impulsivity score, and for the C-BARC energy score. The p-values obtained from the Kruskall-Wallis test are provided in Appendix A. Box diagrams of the serotonin and dopamine levels for the dogs with suspected diagnoses of being ADHD-like (vs the control dogs) and the dogs with scores over and under the median for the different DIAS, Dog ARS, and C-BARQ traits are shown in Figure 1. Regarding the total DIAS score, dogs with lower serotonin and dopamine serum concentrations were more likely to have scores over the median, aside from males and entire animals. In relation to behavioral regulation, it varied with serotonin concentrations in the same direction as the total DIAS score; the model also indicated that the behavioral regulation scores increased with decreased age. Finally, serotonin and dopamine levels were also linked to aggression, aside from gender, dog activity pattern, and number of dogs previously owned (with male dogs being more likely to have DIAS aggression scores over the median, whereas dogs that performed some activity or dogs whose owners who had previously had dogs were more likely to have scores under the median) ( Table 4). No relationship was found between serotonin or dopamine concentration and responsiveness.

Serum Serotonin
Data from the Dog ARS activity-impulsivity indicated that dogs with lower serotonin serum concentrations were more likely to have scores over the median (Table 4). No relationship was found between serotonin or dopamine concentration and Dog ARS inattention.
No relationships were found between serotonin and dopamine serum concentrations and trainability, stranger-directed aggression, owner-directed aggression, dog-directed ag-gression, stranger-directed fear, separation-related problems, and excitability, as measured by the C-BARQ.
However, the probability of a score being over the median for dog-directed fear increased with decreased serotonin concentrations. It also increased in males, entire animals, animals acquired at less than 2 months old, and those whose owners had no dogs previously. With regard to rivalry, the regression model indicated that dogs with lower serotonin had higher C-BARQ scores; higher scores were also found for females and entire animals. Lower serotonin levels likewise increased the scores for non-social fear and touch sensitivity. Touch sensitivity was also higher in dogs acquired at less than 2 months of age, and it increased as weight decreased. Serotonin concentration was also related to attachment (higher scores at lower concentrations). Further, attachment scores were higher in males and with increasing weight and lower in active dogs. Finally, energy scores were higher in animals with lower dopamine levels and in younger dogs (Table 5).

Discussion
In humans, ADHD may be associated with dysregulation of the catecholaminergic and serotonergic systems [30]. Moreover, several studies have shown a decrease in serum serotine in ADHD patients [1,29,30,43]. As in human studies, we found that serotonin and dopamine levels were significantly lower in suspected ADHD-like dogs.
One limitation of the study was the possible ADHD-like diagnosis subjectivity. Because of the lack of a gold standard test, it was based on the behavior signs of being ADHDlike. Nevertheless, three validated questionaries were also used. C-BARQ was used to help in the diagnosis of different behavior problems. Furthermore, the relationships between dopamine, serotonin, and the C-BARQ were not based on the clinical subjective diagnosis. In the same way, the Dog ARS is a validated scale that was developed on the basis of a validated human ADHD questionnaire [8], and the DIAS is a validated owner-based questionnaire used to assess impulsivity [32,41,42]. The use of these different scales allowed us to identify a more objective relationship between the different traits and neurotransmitters.
Regarding the serotonergic system, it was previously thought that dopamine is the key regulator of hyperactivity, but research findings from rodent and hamster models have suggested that serotonin is also involved in hyperactivity [24,[44][45][46]. Furthermore, several human and animal studies have indicated a link between poor impulse control and a low level of the major serotonin metabolite 5-hydroxyindolacetic acid in cerebrospinal fluid [30]. ADHD patients with oppositional defiant disorder (the most frequent coexisting condition in ADHD, characterized by a repeatable pattern of negativistic and defiant behavior) showed lower serum serotonin levels than pure ADHD patients. Furthermore, some tricyclic anti-depressants, as well as monoamine oxidase inhibitors, which are known to increase serotonergic activity, are apparently efficacious in reducing hyperactivity, impulsiveness, and concentration impairment in ADHD children [30,47].
There are few studies related to impulsivity and serotonin in dogs. The DIAS was previously used to assess urine serotonin and dopamine in impulsive and control dogs. Higher DIAS impulsivity scores were found to be significantly correlated with impulsive behavior (reduced tolerance to delay of reinforcement) and lower levels of urinary serotonin and serotonin/dopamine ratios [32]. Further, Peremans et al. [48] supported the role of the serotonergic system in canine impulsive aggression. Similar results were found in the present study in which dogs with higher DIAS scores showed lower serotonin concentrations. Moreover, the other impulsivity-hyperactivity scale used in this study, the Dog ARS, also related impulsivity-hyperactivity scores with lower serotonin levels. In this line, the clinical effectiveness of fluoxetine, a selective serotonin re-uptake inhibitor, in the treatment of ADHD-like dogs [25][26][27] could be linked with the serotonin effect on impulsivity, hyperactivity, and with the intimate interplay between serotonin and dopaminergic neurotransmission [24].
ADHD is frequently associated with comorbid aggression [14], anxiety, and phobias [49][50][51]. Furthermore, ADHD-like dogs have had strong comorbidities with compulsive behaviors, aggressiveness, and fearfulness [52]. A strong relationship between aggression and low serotonin levels has been demonstrated in different species, including dogs [33,34,36]. Similar results were found in the present study where different aggression scales were related to serum serotonin. The Lasso regression indicated that the dogs with lower serotonin levels were more likely to have scores over the median for the DIAS subscale of aggression and for the C-BARQ rivalry behavioral trait. In relation to the above, the aggressive dogs showed low serotonin levels; however, aggression is one of the symptoms and comorbidities of being ADHD-like precisely because of the tendency to behave impulsively, and it is also related to low serotonin levels. Serotonin likely mediates the tendency to behave aggressively and impulsively in ADHD-like dogs.
Animal models consistently demonstrate that brain serotonin systems are critically involved in the response to stressors, as well as in fear and anxiety generation [53][54][55]. Accordingly, lower serotonin concentrations have been found in dogs showing defensive forms of aggression than in other aggression forms [33], and serotoninergic drugs are used to treat fear and phobias in humans [56,57] and dogs [58,59]. Interestingly, in the present paper, significantly lower serotonin levels were found in the dogs with higher C-BARQ scores for fear, anxiety, and touch sensitivity, and touch sensitivity is related to fearful or wary responses to potentially painful or uncomfortable procedures. Moreover, previous research has shown the relationship between serotonin and attachment. In this sense, early maternal deprivation is related to decreasing 5-hydroxyindoleacetic levels in the cerebrospinal fluid and/or through changes in serotonin transporter gene DNA methylation [60][61][62]. Nevertheless, the association between blood serotonin and attachment in dogs was not previous described. Although the importance of serotonin in the social behaviors of different species has already been studied, is still needed to clarify the relationship between blood serotonin and attachment in dogs.
Since approximately 70% of the children and adolescents with ADHD successfully respond to methylphenidate [1,15] and some ADHD-like dogs respond to this treatment as well [16,17], it appears that dopamine has an important role in ADHD. Some genes of the dopaminergic neurotransmitter system were associated with activity, impulsivity, and attention deficits in dogs [19,20,63,64]. Nevertheless, this is the first study that showed a relationship between ADHD-like dogs and serum dopamine. More specifically, it appeared that the dogs with high activity scores were more likely to show lower dopamine levels. Dopamine functions are related to voluntary movement. Therefore, studies on hamsters indicated that Rovorowsky hamsters (which are more hyperactive than Djungarian hamsters) had lower levels of brain dopamine [45].
No relationship was found between serotonin or dopamine concentration and attention deficits. Furthermore, police German Shepherd dogs (possessing polymorphism in exon 3 of the dopamine D4 receptor gene) showed significantly higher scores in the activityimpulsivity component of the dog-ADHD rating scale than did dogs without this allele. This relationship was not found for the attention deficit component [20]. More studies appear to be necessary in order to deepen the relationship between attention deficits and dopamine and serotonin levels, or perhaps the serum levels of dopamine and serotonin may not be good markers for studying attention deficits.
A higher incidence of human ADHD is described in boys (compared to girls) [1,2]. Comparable results were found in our study where males suffered from being ADHD-like significantly more frequently than females, aside from having higher DIAS scores. Previous research has found similar risk factors for dogs [37,65]. In addition, neutering appeared to be an important risk factor for being ADHD-like, and entire animals were more prone to suffering ADHD-like behaviors than castrated dogs and had higher DIAS scores, as previously described [37,65]. Nevertheless, Zink et al. [66] found that neutering could be a risk factor for hyperactivity, whereas Fadel et al. [67] indicated that castrated animals showed higher DIAS scores. Nevertheless, some owners could try to find a solution for dealing with the condition of their dog by neutering. Hence, the role of sex hormones in ADHD-like behaviors should be studied in more detail.
Interestingly, age was also a significant contributor to the prevalence of ADHD-like symptoms. Furthermore, the DIAS behavior regulation score increased with decreasing age. Other studies have found that hyperactive, impulsive, and inattentive behaviors were much more prevalent in young dogs [9,19,41,65]. Likewise, in humans, ADHD is age-dependent, being more prevalent in youths [4], in spite of the possibility that the condition can remain present in adult humans. Young dogs may appear more hyperactive to inexperienced owners; however, the possible effect of age was used as a control variable in the multivariate analysis.
Dogs that performed some activity and dogs from owners who had previously had dogs were more likely to have scores below the median for DIAS aggression. Other studies observed associations between an owner's dog experience and a dog's aggressive behavior [52,[54][55][56][57][68][69][70]. It is possible that experienced owners are more aware of the importance of socialization. Previous experience can also help owners to identify a problem at an early stage, when the problem can be treated more efficiently. Furthermore, if the owners had problems with their first dogs, they may be more careful when choosing a new dog [52]. On the other hand, they could have a comparison point if the previous dog was behaving normally. Hence, they might be more aware that something is not normal and refer to it earlier. In addition, they might have more experience in dog training, which could be a part of the improvement in their dog.
Activity patterns had a relationship with the presence of aggression. The complete underlying mechanism behind the effect of exercise in aggression is not clear, but it is known that exercise increases serotonin production both in animals and humans, thus functioning as an antidepressant [71], and it was also related to reduced anxiety [72] and aggression in dogs [70]. In addition, in ADHD-like cases, owners might have difficulties in walking or exercising their dogs. Consequently, the quantity of exercise might be a consequence of the dog being challenging for months, rather than a correlation or a cause.
Finally, the paper highlights the importance of pairing assessments by veterinary specialists using validated tests with avoiding assessments by other professionals. Further, it would be interesting to assess the role of genetics and neurotransmitters in ADHD-like dogs in future studies.

Conclusions
Our data suggest a strong relationship between serotonin and dopamine and ADHDlike dogs. Furthermore, lower serotonin and dopamine serum concentrations were related to being ADHD-like. Moreover, serotonin and dopamine levels were also linked to aggres-sion, activity, and impulsivity. Decreased serotonin concentrations were also related to fear, attachment problems, and touch sensitivity.
No relationship was found between serotonin and dopamine serum concentrations and attention deficits.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ani13061037/s1, Table S1: Study data set.  Informed Consent Statement: Informed consent was obtained from all the dogs' owners involved in the study.

Data Availability Statement:
The data presented in this study are available on request from the first author/corresponding authors.

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. The p-values obtained from the Kruskal-Wallis tests used to compare the mean serotonin and dopamine serum concentrations (ng/mL) in the dogs with a suspected diagnosis of being ADHDlike (vs. not ADHD-like dogs) and the dogs with scores over and under the median for the different DIAS, Dog ARS, and C-BARQ traits are shown in Table A1.